模型价格高速 ↑文档个人中心轮盘Stealth Warden
ENRU中文हिन्दी
最佳 AI 网关

Gwarden AI
93 models

93 个模型，覆盖代码、智能体与各类任务 - 一个简单 API 即可使用。
在 Telegram 获取密钥，免费 7 天试用开始。

个人中心 文档 幸运轮盘 网关不工作？
93 models OpenAI + Anthropic API Claude Code · Codex · opencode · aider 上下文最高 1M 免费开始 - 7 天
模型

模型系列

Claude Fable 5.1

Anthropic
API

Fable 5 的增量更新：更精准的指令遵循、长程编程与工具调用。

$3.50 in$13.00 out1M ctx65.7% Q
状态 98.89%延迟 4.88sTPS 1.2t/s

Claude Opus 5.5

Anthropic
General

高级阵容 — 属于 General 套餐。

$55/mo
获取套餐

GPT Astra 6

OpenAI
General

高级阵容 — 属于 General 套餐。

$55/mo
获取套餐

Claude Opus 5

Anthropic
API

最强的一代 Opus。为复杂长任务与深度研究而生。

$2.00 in$10.00 out1M ctx63.1% Q
状态 92.22%延迟 47.44sTPS 14.0t/s

GPT 6.1 Sol

OpenAI
General

高级阵容 — 属于 General 套餐。

$55/mo
获取套餐

Claude Fable 5

Anthropic
API

Anthropic 2026 旗舰。Claude 家族最强的编程模型：100万上下文、视觉输入。

$3.50 in$13.00 out1M ctx62.1% Q
状态 91.11%延迟 30.46sTPS 15.7t/s

GPT 6 Sol

OpenAI
General

高级阵容 — 属于 General 套餐。

$55/mo
获取套餐

Stealth Warden

Gwarden
API

我们自有的前沿模型，运行在专属通道上。Opus 级回答，1M 上下文，独立余额，与主网关不共享算力。

$0.06 in$0.17 out1M ctx62.0% Q
状态 100.00%延迟 -TPS -

Claude Sonnet 5.5

Anthropic
General

高级阵容 — 属于 General 套餐。

$55/mo
获取套餐

GPT-5.6 Sol

OpenAI
API

GPT-5.6 家族的前沿档。为高难度编程与代理任务提供顶级推理。

$3.50 in$18.00 out1M ctx60.9% Q
状态 94.44%延迟 22.65sTPS 16.9t/s

Grok 4.6

xAI
API

xAI 前沿模型。强推理与编程，具备实时知识。

$1.70 in$5.00 out1M ctx60.9% Q
状态 93.33%延迟 27.54sTPS 11.6t/s

Kimi K3

Moonshot AI
API

Moonshot AI 旗舰开源模型。顶级编程与代理能力，100万 上下文。

$2.60 in$13.00 out1M ctx59.7% Q
状态 98.89%延迟 28.34sTPS 16.7t/s

Hy4

Tencent
API

腾讯混元旗舰层。

$1.60 in$7.50 out1M ctx59.5% Q
状态 97.78%延迟 13.50sTPS 14.6t/s

GLM 5.3

Z.ai
API

Z.ai 旗舰模型。100万 token 上下文、推理始终开启，代码与代理任务的顶尖选择。

$1.20 in$2.40 out1M ctx59.5% Q
状态 97.78%延迟 5.72sTPS -

Gemini 3.8 Flash

Google
API

Google 快速版 Gemini：低延迟、多模态、100万上下文。

$1.00 in$4.50 out1M ctx58.7% Q
状态 98.89%延迟 21.80sTPS 18.3t/s

Qwen3.8 Max

Alibaba
API

阿里最强 Qwen。编程与多语言表现出色，100万上下文。

$1.70 in$5.20 out1M ctx58.1% Q
状态 100.00%延迟 1.55sTPS 1.9t/s

Gwarden AI

Gwarden
API

Gwarden AI——友好的编程助手，100万上下文。任意 $2 一次性网关购买即可解锁，永久拥有。

$0.70 in$1.50 out1M ctx57.9% Q
状态 100.00%延迟 12.54sTPS 1.5t/s

Qwen3.8 2.4T A95B

Alibaba
API

开源 Qwen MoE：总参 2.4T、激活 95B。开源价格的前沿质量。

$0.80 in$1.70 out1M ctx57.7% Q
状态 85.56%延迟 20.93sTPS 14.0t/s

Claude Haiku 5.5

Anthropic
General

高级阵容 — 属于 General 套餐。

$55/mo
获取套餐

GPT-5.6 Terra

OpenAI
API

GPT-5.6 家族的均衡中档：以合理价格获得前沿质量。

$1.70 in$10.00 out1M ctx56.5% Q
状态 91.11%延迟 24.85sTPS 17.3t/s

Claude Sonnet 5

Anthropic
API

Claude 家族的均衡主力：快速、编程强、100万上下文。

$1.70 in$8.50 out1M ctx56.4% Q
状态 92.22%延迟 25.59sTPS 20.6t/s

Hy3

Tencent
API

腾讯混元均衡层。

$0.90 in$3.20 out1M ctx56.0% Q
状态 98.89%延迟 13.33sTPS 14.7t/s

GPT-5.5

OpenAI
API

OpenAI 上一代旗舰，依然是可靠的生产主力。

$4.50 in$27.00 out1M ctx55.8% Q
状态 92.22%延迟 28.14sTPS 12.2t/s

Muse Spark 1.3

Meta
API

Meta 的快速 contributor 级模型。100万上下文，聊天、编程与创作皆宜。

$0.40 in$1.60 out1M ctx55.1% Q
状态 100.00%延迟 -TPS -

DeepSeek V4.1 Flash

DeepSeek
API

DeepSeek 快速层，质量接近旗舰。

$0.35 in$1.10 out1M ctx55.0% Q
状态 100.00%延迟 -TPS -

DeepSeek V4 Pro

DeepSeek
API

DeepSeek V4 旗舰：推理优先的编程模型，基准成绩顶尖。

$0.80 in$2.20 out1M ctx54.5% Q
状态 100.00%延迟 -TPS -

GLM-5.2

Z.ai
API

GLM 旗舰的上一代。性价比很高的编程模型。

$1.20 in$3.90 out1M ctx53.5% Q
状态 97.78%延迟 20.83sTPS 21.6t/s

GPT-5.6 Luna

OpenAI
API

GPT-5.6 家族的轻量快速档。超低价的日常主力。

$0.18 in$1.05 out1M ctx52.1% Q
状态 100.00%延迟 -TPS -

Qwen3.8 27B

Alibaba
API

紧凑开源 Qwen（27B）：快速、便宜且出奇能干。

$0.22 in$0.80 out1M ctx52.0% Q
状态 100.00%延迟 -TPS -

DeepSeek V4 Flash

DeepSeek
API

DeepSeek V4 快速版：接近旗舰的质量，延迟与成本更低。

$0.30 in$0.90 out1M ctx52.0% Q
状态 100.00%延迟 2.69sTPS 1.6t/s

Gemini 3.1 Pro

Google
API

Google Gemini Pro 档：100万上下文，复杂推理与多模态。

$1.75 in$10.50 out1M ctx51.2% Q
状态 93.33%延迟 29.27sTPS 8.5t/s

MiMo V2.6-Pro

Xiaomi
API

小米 MiMo V2.6-Pro 推理模型。

$0.70 in$2.60 out1M ctx51.0% Q
状态 94.44%延迟 18.32sTPS 18.6t/s

GPT-OSS 120B

OpenAI
API

OpenAI 开源权重模型 GPT-OSS（MoE 120B，约 5B 激活）。接近旗舰的质量，开源价格。

$0.45 in$1.80 out1M ctx50.6% Q
状态 96.67%延迟 17.65sTPS 14.3t/s

Agnes 2.5 Flash

Sapiens AI
API

Sapiens AI 的 Flash 代快速模型。低延迟、100万上下文、推理始终开启。

$0.35 in$1.40 out1M ctx49.8% Q
状态 82.22%延迟 42.62sTPS 11.9t/s

MiniMax M3

MiniMax
API

MiniMax M3：面向代理与日常编程的快速长上下文模型。

$0.27 in$1.05 out1M ctx47.5% Q
状态 83.33%延迟 16.64sTPS 13.8t/s

Mimo 2.5 Pro

Xiaomi
API

小米 MiMo Pro：高性价比的推理模型，编程能力扎实。

$0.60 in$2.20 out1M ctx47.5% Q
状态 96.67%延迟 14.20sTPS 11.1t/s

DeepSeek R1 0528 Qwen3 8B

DeepSeek
API

DeepSeek R1 (0528) 蒸馏到 Qwen3-8B 的推理模型。小巧快速，数学与逻辑见长。

$0.12 in$0.45 out1M ctx44.2% Q
状态 80.65%延迟 17.47sTPS 20.4t/s

Claude Opus 4.8

Anthropic
API

上一代 Opus，依然是生产环境的主力。

$2.00 in$10.00 out1M ctx
状态 94.44%延迟 24.72sTPS 14.9t/s

GPT Image 2

OpenAI
API

GPT Image 2：照片级图像生成。每次固定 $1.00，不按 token 计费——一次调用，一张图。

$1.00 / 次
ModelInput / 1MOutput / 1MCache / 1MQuality
claude-fable-5-11M / 128k out
Details

Fable 5 的增量更新：更精准的指令遵循、长程编程与工具调用。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $3.50 in / $13.00 out / $4.00 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-fable-5-1", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$3.50$13.00$4.0065.7%
claude-opus-5.51M / 64k out
Details

高级阵容 — 属于 General 套餐。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $4.50 in / $22.00 out / $0.45 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-opus-5.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$4.50$22.00$0.4564.0%
gpt-astra-61M / 128k out
Details

高级阵容 — 属于 General 套餐。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $5.00 in / $24.00 out / $0.50 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-astra-6", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$5.00$24.00$0.5063.5%
claude-opus-51M / 64k out
Details

最强的一代 Opus。为复杂长任务与深度研究而生。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $2.00 in / $10.00 out / $2.00 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-opus-5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$2.00$10.00$2.0063.1%
gpt-6.1-sol1M / 128k out
Details

高级阵容 — 属于 General 套餐。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $4.20 in / $20.00 out / $0.42 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-6.1-sol", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$4.20$20.00$0.4262.5%
claude-fable-51M / 128k out
Details

Anthropic 2026 旗舰。Claude 家族最强的编程模型：100万上下文、视觉输入。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $3.50 in / $13.00 out / $4.00 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-fable-5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$3.50$13.00$4.0062.1%
gpt-6-sol1M / 128k out
Details

高级阵容 — 属于 General 套餐。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $3.80 in / $18.00 out / $0.38 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-6-sol", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$3.80$18.00$0.3862.0%
stealth-warden1M / 128k out
Details

我们自有的前沿模型，运行在专属通道上。Opus 级回答，1M 上下文，独立余额，与主网关不共享算力。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Gwarden  Price: $0.06 in / $0.17 out / $0.01 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "stealth-warden", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.06$0.17$0.0162.0%
claude-sonnet-5.51M / 64k out
Details

高级阵容 — 属于 General 套餐。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $2.60 in / $12.00 out / $0.26 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-5.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$2.60$12.00$0.2661.0%
gpt-5.6-sol1M / 128k out
Details

GPT-5.6 家族的前沿档。为高难度编程与代理任务提供顶级推理。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $3.50 in / $18.00 out / $0.35 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5.6-sol", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$3.50$18.00$0.3560.9%
grok-4.61M / 64k out
Details

xAI 前沿模型。强推理与编程，具备实时知识。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): xAI  Price: $1.70 in / $5.00 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "grok-4.6", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.00$0.4060.9%
kimi-k31M / 64k out
Details

Moonshot AI 旗舰开源模型。顶级编程与代理能力，100万 上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Moonshot AI  Price: $2.60 in / $13.00 out / $0.55 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k3", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$2.60$13.00$0.5559.7%
glm-5.31M / 128k out
Details

Z.ai 旗舰模型。100万 token 上下文、推理始终开启，代码与代理任务的顶尖选择。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $1.20 in / $2.40 out / $0.26 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.3", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.20$2.40$0.2659.5%
hy41M / 128k out
Details

腾讯混元旗舰层。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Tencent  Price: $1.60 in / $7.50 out / $0.30 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "hy4", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.60$7.50$0.3059.5%
gemini-3.8-flash1M / 64k out
Details

Google 快速版 Gemini：低延迟、多模态、100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Google  Price: $1.00 in / $4.50 out / $0.25 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3.8-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.00$4.50$0.2558.7%
qwen3.8-max1M / 64k out
Details

阿里最强 Qwen。编程与多语言表现出色，100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $1.70 in / $5.20 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.8-max", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.20$0.4058.1%
gwarden-ai1M / 64k out
Details

Gwarden AI——友好的编程助手，100万上下文。任意 $2 一次性网关购买即可解锁，永久拥有。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Gwarden  Price: $0.70 in / $1.50 out / $0.09 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gwarden-ai", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.70$1.50$0.0957.9%
qwen3.8-2.4t-a95b1M / 64k out
Details

开源 Qwen MoE：总参 2.4T、激活 95B。开源价格的前沿质量。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $0.80 in / $1.70 out / $0.15 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.8-2.4t-a95b", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.80$1.70$0.1557.7%
claude-haiku-5.51M / 64k out
Details

高级阵容 — 属于 General 套餐。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $0.80 in / $3.50 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-haiku-5.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.80$3.50$0.0857.5%
gpt-5.6-terra1M / 128k out
Details

GPT-5.6 家族的均衡中档：以合理价格获得前沿质量。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $1.70 in / $10.00 out / $0.18 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5.6-terra", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$10.00$0.1856.5%
claude-sonnet-51M / 64k out
Details

Claude 家族的均衡主力：快速、编程强、100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $1.70 in / $8.50 out / $0.17 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$8.50$0.1756.4%
Qwen3.8-Flash-Next1M / 64k out
Details

Text model served through the Gwarden gateway with OpenAI-compatible API.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $0.15 in / $0.50 out / $0.03 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "Qwen3.8-Flash-Next", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.15$0.50$0.0356.0%
hy31M / 64k out
Details

腾讯混元均衡层。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Tencent  Price: $0.90 in / $3.20 out / $0.20 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "hy3", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.90$3.20$0.2056.0%
gpt-5.51M / 128k out
Details

OpenAI 上一代旗舰，依然是可靠的生产主力。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $4.50 in / $27.00 out / $0.45 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$4.50$27.00$0.4555.8%
muse-spark-1.3-contributor1M / 64k out
Details

Meta 的快速 contributor 级模型。100万上下文，聊天、编程与创作皆宜。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Meta  Price: $0.40 in / $1.60 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "muse-spark-1.3-contributor", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.40$1.60$0.0855.1%
deepseek-v4.1-flash1M / 64k out
Details

DeepSeek 快速层，质量接近旗舰。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.35 in / $1.10 out / $0.09 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4.1-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.35$1.10$0.0955.0%
deepseek-v4-pro1M / 64k out
Details

DeepSeek V4 旗舰：推理优先的编程模型，基准成绩顶尖。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.80 in / $2.20 out / $0.20 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-pro", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.80$2.20$0.2054.5%
glm-5.21M / 128k out
Details

GLM 旗舰的上一代。性价比很高的编程模型。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $1.20 in / $3.90 out / $0.25 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.2", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.20$3.90$0.2553.5%
gpt-5.6-luna1M / 128k out
Details

GPT-5.6 家族的轻量快速档。超低价的日常主力。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $0.18 in / $1.05 out / $0.01 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5.6-luna", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.18$1.05$0.0152.1%
deepseek-v4-flash1M / 32k out
Details

DeepSeek V4 快速版：接近旗舰的质量，延迟与成本更低。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.30 in / $0.90 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.30$0.90$0.0852.0%
qwen3.8-27b1M / 32k out
Details

紧凑开源 Qwen（27B）：快速、便宜且出奇能干。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $0.22 in / $0.80 out / $0.05 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.8-27b", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.22$0.80$0.0552.0%
gemini-3.1-pro1M / 64k out
Details

Google Gemini Pro 档：100万上下文，复杂推理与多模态。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Google  Price: $1.75 in / $10.50 out / $0.45 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3.1-pro", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.75$10.50$0.4551.2%
mimo-v2.6-pro1M / 64k out
Details

小米 MiMo V2.6-Pro 推理模型。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Xiaomi  Price: $0.70 in / $2.60 out / $0.18 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "mimo-v2.6-pro", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.70$2.60$0.1851.0%
gpt-oss-120b1M / 64k out
Details

OpenAI 开源权重模型 GPT-OSS（MoE 120B，约 5B 激活）。接近旗舰的质量，开源价格。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $0.45 in / $1.80 out / $0.09 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-oss-120b", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.45$1.80$0.0950.6%
agnes-2.5-flash1M / 64k out
Details

Sapiens AI 的 Flash 代快速模型。低延迟、100万上下文、推理始终开启。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Sapiens AI  Price: $0.35 in / $1.40 out / $0.07 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "agnes-2.5-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.35$1.40$0.0749.8%
mimo-v2.5-pro1M / 64k out
Details

小米 MiMo Pro：高性价比的推理模型，编程能力扎实。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Xiaomi  Price: $0.60 in / $2.20 out / $0.15 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "mimo-v2.5-pro", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.60$2.20$0.1547.5%
minimax-m31M / 64k out
Details

MiniMax M3：面向代理与日常编程的快速长上下文模型。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): MiniMax  Price: $0.27 in / $1.05 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "minimax-m3", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.27$1.05$0.0847.5%
deepseek-r1-0528-qwen3-8b1M / 32k out
Details

DeepSeek R1 (0528) 蒸馏到 Qwen3-8B 的推理模型。小巧快速，数学与逻辑见长。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.12 in / $0.45 out / $0.03 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-r1-0528-qwen3-8b", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.12$0.45$0.0344.2%
Ling-3.1-flash256k / 64k out
Details

蚂蚁 Ling 3.1 Flash：混合推理 MoE，面向智能体与长周期开发。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Ant Group  Price: $0.40 in / $1.60 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "Ling-3.1-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.40$1.60$0.08-
agnes-3.0-flash1M / 64k out
Details

Sapiens AI's newest Flash-generation model. Low latency, 1M context, always-on reasoning.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Sapiens AI  Price: $0.50 in / $2.00 out / $0.10 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "agnes-3.0-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.50$2.00$0.10-
azure-tts0k / 0k out
Details

Microsoft Azure neural text-to-speech: 300+ voices, natural prosody. Billed per 1M characters.

How it works: Speech model: sends your text and returns playable audio bytes directly. Billed per 1M input characters. Works with POST /v1/audio/speech.

Endpoint: POST /v1/audio/speech  Vendor (upstream): Microsoft Azure  Price: $16 /1M chars

API request example:

curl https://gwarden.su/v1/audio/speech \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "azure-tts", "input": "Hello, this is a test.", "voice": "en-US-AvaNeural", "response_format": "mp3"}' --output speech.mp3
$0.15$0.50$0.03-
bge-large-en-v1.58k / 0k out
Details

BAAI BGE embeddings: multilingual text vectors for search and RAG. Billed per 1M tokens.

How it works: Embedding model: turns text into numeric vectors for search, clustering and RAG. Billed per 1M input tokens. Works with POST /v1/embeddings.

Endpoint: POST /v1/embeddings  Vendor (upstream): BAAI  Price: $0.05 /1M tokens

API request example:

curl https://gwarden.su/v1/embeddings \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "bge-large-en-v1.5", "input": "Hello world"}'
$0.05$0.00$0.00-
bge-large-zh-v1.58k / 0k out
Details

BAAI BGE embeddings: multilingual text vectors for search and RAG. Billed per 1M tokens.

How it works: Embedding model: turns text into numeric vectors for search, clustering and RAG. Billed per 1M input tokens. Works with POST /v1/embeddings.

Endpoint: POST /v1/embeddings  Vendor (upstream): BAAI  Price: $0.05 /1M tokens

API request example:

curl https://gwarden.su/v1/embeddings \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "bge-large-zh-v1.5", "input": "Hello world"}'
$0.05$0.00$0.00-
bge-m38k / 0k out
Details

BAAI BGE embeddings: multilingual text vectors for search and RAG. Billed per 1M tokens.

How it works: Embedding model: turns text into numeric vectors for search, clustering and RAG. Billed per 1M input tokens. Works with POST /v1/embeddings.

Endpoint: POST /v1/embeddings  Vendor (upstream): BAAI  Price: $0.05 /1M tokens

API request example:

curl https://gwarden.su/v1/embeddings \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "bge-m3", "input": "Hello world"}'
$0.05$0.00$0.00-
bge-reranker-v2-m38k / 0k out
Details

BAAI BGE embeddings: multilingual text vectors for search and RAG. Billed per 1M tokens.

How it works: Embedding model: turns text into numeric vectors for search, clustering and RAG. Billed per 1M input tokens. Works with POST /v1/embeddings.

Endpoint: POST /v1/embeddings  Vendor (upstream): BAAI  Price: $0.1 /1M tokens

API request example:

curl https://gwarden.su/v1/embeddings \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "bge-reranker-v2-m3", "input": "Hello world"}'
$0.10$0.00$0.00-
claude-opus-4-81M / 64k out
Details

上一代 Opus，依然是生产环境的主力。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Anthropic  Price: $2.00 in / $10.00 out / $2.00 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-opus-4-8", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$2.00$10.00$2.00-
deepseek-ocr1M / 32k out
Details

DeepSeek OCR: reads text, tables and formulas off images and documents through chat with vision.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.50 in / $1.50 out / $0.12 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-ocr", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.50$1.50$0.12-
deepseek-v4-flash-0731256k / 32k out
Details

DeepSeek V4 快速版：接近旗舰的质量，延迟与成本更低。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.30 in / $0.90 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-flash-0731", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.30$0.90$0.08-
deepseek-v4-flash-0731-3256k / 32k out
Details

DeepSeek V4 快速版：接近旗舰的质量，延迟与成本更低。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.15 in / $0.50 out / $0.04 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-flash-0731-3", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.15$0.50$0.04-
deepseek-v4-flash-vision-exp256k / 32k out
Details

DeepSeek V4 快速版：接近旗舰的质量，延迟与成本更低。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.20 in / $0.60 out / $0.05 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-flash-vision-exp", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.20$0.60$0.05-
deepseek-v4-pro-08131M / 64k out
Details

DeepSeek V4 旗舰：推理优先的编程模型，基准成绩顶尖。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): DeepSeek  Price: $0.80 in / $2.20 out / $0.20 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "deepseek-v4-pro-0813", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.80$2.20$0.20-
doubao-seed-2-1-lite-260915256k / 32k out
Details

字节 Doubao Seed 2.1 Lite：日常对话的快速轻量模型。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): ByteDance  Price: $0.20 in / $0.70 out / $0.05 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "doubao-seed-2-1-lite-260915", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.20$0.70$0.05-
doubao-seed-2-1-pro256k / 64k out
Details

ByteDance Doubao Seed 2.1: strong Chinese/English model for chat and agents.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): ByteDance  Price: $0.50 in / $2.00 out / $0.12 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "doubao-seed-2-1-pro", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.50$2.00$0.12-
doubao-seed-2-1-turbo256k / 64k out
Details

ByteDance Doubao Seed 2.1: strong Chinese/English model for chat and agents.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): ByteDance  Price: $0.30 in / $1.20 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "doubao-seed-2-1-turbo", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.30$1.20$0.08-
gemini-3.1-pro-preview1M / 64k out
Details

Google Gemini Pro 档：100万上下文，复杂推理与多模态。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Google  Price: $1.75 in / $10.50 out / $0.45 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3.1-pro-preview", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.75$10.50$0.45-
gemini-3.7-flash1M / 64k out
Details

Google's new fast Gemini: low latency, multimodal, 1M context.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Google  Price: $1.00 in / $4.50 out / $0.25 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-3.7-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.00$4.50$0.25-
gemini-flash1M / 32k out
Details

Google 快速版 Gemini：低延迟、多模态、100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Google  Price: $0.30 in / $1.20 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.30$1.20$0.08-
glm-4-9b-0414128k / 32k out
Details

Compact open GLM (9B): fast and cheap for everyday tasks.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $0.15 in / $0.60 out / $0.03 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-4-9b-0414", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.15$0.60$0.03-
glm-5200k / 64k out
Details

智谱旗舰 GLM-5：顶级编程与智能体工程。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $0.90 in / $2.90 out / $0.18 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.90$2.90$0.18-
glm-5.11M / 64k out
Details

Z.ai's GLM 5.1: previous flagship tier, strong coder for its price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $1.00 in / $3.00 out / $0.20 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.1", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.00$3.00$0.20-
glm-5.3-flash256k / 32k out
Details

Fast lightweight GLM 5.3 variant: near-flagship quality, minimal latency and cost.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $0.40 in / $1.40 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.3-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.40$1.40$0.08-
glm-5.3-flash-2256k / 32k out
Details

Fast lightweight GLM 5.3 variant: near-flagship quality, minimal latency and cost.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $0.10 in / $0.40 out / $0.02 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.3-flash-2", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.10$0.40$0.02-
glm-5.3-flashx256k / 32k out
Details

Fast lightweight GLM 5.3 variant: near-flagship quality, minimal latency and cost.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $0.40 in / $1.40 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.3-flashx", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.40$1.40$0.08-
glm-5.3-prime1M / 64k out
Details

Z.ai 旗舰模型。100万 token 上下文、推理始终开启，代码与代理任务的顶尖选择。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $1.50 in / $3.00 out / $0.30 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-5.3-prime", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.50$3.00$0.30-
glm-z1-9b-0414128k / 32k out
Details

Compact open GLM (9B): fast and cheap for everyday tasks.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Z.ai  Price: $0.15 in / $0.60 out / $0.03 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "glm-z1-9b-0414", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.15$0.60$0.03-
gpt-image-2图像生成
Details

GPT Image 2：照片级图像生成。每次固定 $1.00，不按 token 计费——一次调用，一张图。

How it works: Image model: generates pictures from a text prompt. Billed flat per image, no token maths - one call, one picture.

Endpoint: POST /v1/images/generations  Vendor (upstream): OpenAI  Price: $1.00 / 次

API request example:

curl https://gwarden.su/v1/images/generations \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-image-2", "prompt": "A fox in a neon city"}'
$1.00 / 次-
gpt-oss-20b128k / 32k out
Details

OpenAI's compact open-weight GPT-OSS (20B). Fast, cheap, open price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): OpenAI  Price: $0.15 in / $0.60 out / $0.03 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-oss-20b", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.15$0.60$0.03-
grok-4.51M / 64k out
Details

xAI's newest Grok generation. Strong reasoning and coding with real-time knowledge.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): xAI  Price: $1.70 in / $5.00 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "grok-4.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.00$0.40-
grok-4.71M / 64k out
Details

xAI's newest Grok generation. Strong reasoning and coding with real-time knowledge.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): xAI  Price: $1.70 in / $5.00 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "grok-4.7", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.00$0.40-
grok-4.7-build-fast1M / 64k out
Details

xAI's newest Grok generation. Strong reasoning and coding with real-time knowledge.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): xAI  Price: $1.70 in / $5.00 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "grok-4.7-build-fast", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.00$0.40-
grok-build-0.11M / 64k out
Details

xAI's newest Grok generation. Strong reasoning and coding with real-time knowledge.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): xAI  Price: $1.00 in / $3.00 out / $0.25 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "grok-build-0.1", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.00$3.00$0.25-
hy4-preview1M / 64k out
Details

腾讯混元旗舰层。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Tencent  Price: $1.60 in / $7.50 out / $0.30 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "hy4-preview", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.60$7.50$0.30-
indextts-20k / 0k out
Details

IndexTTS2: voice cloning from a reference audio clip. Billed per 1M characters.

How it works: Speech model: sends your text and returns playable audio bytes directly. Billed per 1M input characters. Works with POST /v1/audio/speech.

Endpoint: POST /v1/audio/speech  Vendor (upstream): IndexTTS  Price: $8 /1M chars

API request example:

curl https://gwarden.su/v1/audio/speech \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "indextts-2", "input": "Hello, this is a test.", "voice": "en-US-AvaNeural", "response_format": "mp3"}' --output speech.mp3
$0.15$0.50$0.03-
jev-1.13按次固定计费
Details

Jev 1.13 by TypeSafe AI: a decision model, not a chatbot. It evaluates your data against typed questions (yes/no, choice, score) and returns probabilities with confidence. Flat price per request.

How it works: Decision model: does NOT generate text. You send a state block plus questions (noul = yes/no 0..1, choice = pick one option, score = grade on your scale) and get structured answers with probabilities. Billed flat per request, no matter how many questions are in one call.

Endpoint: POST /v1/systemone  Vendor (upstream): TypeSafe AI  Price: $0.0010 / 次

API request example:

curl https://gwarden.su/v1/systemone \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "jev-1.13", "state": {"ticket": "Card was charged twice, need urgent help"}, "questions": {"needs_human": {"type": "noul", "instructions": "Does this need a human now?"}, "route": {"type": "choice", "instructions": "Which team owns this?", "criteria": {"billing": "payments and refunds", "support": "product usage"}}, "urgency": {"type": "score", "instructions": "How urgent is this?", "criteria": ["can wait", "handle soon", "handle now"]}}}'
$0.0010 / 次-
kimi-k2.5256k / 64k out
Details

Moonshot AI's K2 line: open-weight agents and coders at a mid-tier price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Moonshot AI  Price: $0.60 in / $2.50 out / $0.15 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k2.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.60$2.50$0.15-
kimi-k2.6256k / 64k out
Details

Moonshot AI's K2 line: open-weight agents and coders at a mid-tier price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Moonshot AI  Price: $0.60 in / $2.50 out / $0.15 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k2.6", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.60$2.50$0.15-
kimi-k2.7-code256k / 64k out
Details

Moonshot AI's K2 line: open-weight agents and coders at a mid-tier price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Moonshot AI  Price: $0.60 in / $2.50 out / $0.15 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k2.7-code", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.60$2.50$0.15-
kimi-k2.8-preview256k / 64k out
Details

Moonshot AI's K2 line: open-weight agents and coders at a mid-tier price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Moonshot AI  Price: $0.60 in / $2.50 out / $0.15 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "kimi-k2.8-preview", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.60$2.50$0.15-
laguna-xs-2.1128k / 16k out
Details

Poolside's compact Laguna XS coder: tiny, fast and cheap.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Poolside  Price: $0.10 in / $0.30 out / $0.02 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "laguna-xs-2.1", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.10$0.30$0.02-
mimo-v2.5256k / 32k out
Details

小米 MiMo Pro：高性价比的推理模型，编程能力扎实。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Xiaomi  Price: $0.50 in / $1.80 out / $0.12 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "mimo-v2.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.50$1.80$0.12-
mimo-v2.6-flash256k / 32k out
Details

Xiaomi's fast MiMo flash tier: budget reasoning with solid coding chops.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Xiaomi  Price: $0.30 in / $1.10 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "mimo-v2.6-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.30$1.10$0.08-
mineru2.5-pro按次固定计费
Details

MinerU document parsing: PDF pages into structured markdown. Flat price per request.

How it works: Document parser: turns PDF pages and scans into structured text. Flat price per request. Works with POST /v1/ocr.

Endpoint: POST /v1/ocr  Vendor (upstream): OpenDataLab  Price: $0.01 / 次

API request example:

curl https://gwarden.su/v1/ocr \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "mineru2.5-pro", "document": "https://example.com/file.pdf"}'
$0.01 / 次-
minimax-h3video generation
Details

MiniMax H3 video generation: 1-15s MP4 from text, images or audio reference. Billed per output second.

How it works: Video model: you send a prompt (plus optional reference images/audio) and get a task id; poll the task until it completes, then download the MP4. Billed per output second by resolution.

Endpoint: POST /v1/videos  Vendor (upstream): MiniMax  Price: $0.1/s 1088p / $0.13/s 1440p / $0.05/s 480p / $0.08/s 768p

API request example:

curl https://gwarden.su/v1/videos \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "minimax-h3", "prompt": "A fox crossing a sunlit meadow, slow tracking shot", "seconds": 5, "resolution": "768p", "aspect_ratio": "16:9"}'
$0.05–$0.13 per second-
minimax-m2.5256k / 64k out
Details

MiniMax M2.5：面向代理与日常编程的快速长上下文模型。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): MiniMax  Price: $0.22 in / $0.90 out / $0.07 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "minimax-m2.5", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.22$0.90$0.07-
minimax-m2.7256k / 64k out
Details

MiniMax M2.7: agent-tuned chat model, fast long-context everyday workhorse.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): MiniMax  Price: $0.30 in / $1.20 out / $0.08 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "minimax-m2.7", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.30$1.20$0.08-
minimax-m2.7-highspeed256k / 64k out
Details

MiniMax M2.7: agent-tuned chat model, fast long-context everyday workhorse.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): MiniMax  Price: $0.45 in / $1.80 out / $0.10 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "minimax-m2.7-highspeed", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.45$1.80$0.10-
qwen3.7-flash256k / 32k out
Details

Fast Qwen3.8 variant: near-flagship quality at flash latency and price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $0.35 in / $1.10 out / $0.09 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.7-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.35$1.10$0.09-
qwen3.7-max1M / 64k out
Details

阿里最强 Qwen。编程与多语言表现出色，100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $1.70 in / $5.20 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.7-max", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.20$0.40-
qwen3.7-plus1M / 64k out
Details

阿里最强 Qwen。编程与多语言表现出色，100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $1.00 in / $3.00 out / $0.25 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.7-plus", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.00$3.00$0.25-
qwen3.8-flash256k / 32k out
Details

Fast Qwen3.8 variant: near-flagship quality at flash latency and price.

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $0.35 in / $1.10 out / $0.09 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.8-flash", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.35$1.10$0.09-
qwen3.8-max-09021M / 64k out
Details

阿里最强 Qwen。编程与多语言表现出色，100万上下文。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): Alibaba  Price: $1.70 in / $5.20 out / $0.40 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "qwen3.8-max-0902", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$1.70$5.20$0.40-
sensenova-6.8-flash-lite256k / 32k out
Details

商汤 SenseNova 6.8 Flash Lite：日常对话的快速轻量模型。

How it works: Chat model: sends your messages to the engine and returns generated text. Billed per 1M tokens (input / output / cache). Works with POST /v1/chat/completions, POST /v1/responses and POST /v1/messages.

Endpoint: POST /v1/chat/completions  Vendor (upstream): SenseTime  Price: $0.25 in / $0.90 out / $0.06 cache per 1M

API request example:

curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "sensenova-6.8-flash-lite", "messages": [{"role": "user", "content": "Hello"}], "stream": false}'
$0.25$0.90$0.06-
unreal-speech-v80k / 0k out
Details

UnrealSpeech V8: ultra-cheap natural narration with word timestamps. Billed per 1M characters.

How it works: Speech model: sends your text and returns playable audio bytes directly. Billed per 1M input characters. Works with POST /v1/audio/speech.

Endpoint: POST /v1/audio/speech  Vendor (upstream): UnrealSpeech  Price: $10 /1M chars

API request example:

curl https://gwarden.su/v1/audio/speech \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "unreal-speech-v8", "input": "Hello, this is a test.", "voice": "en-US-AvaNeural", "response_format": "mp3"}' --output speech.mp3
$0.15$0.50$0.03-
Gateway's

优先 API 端点

专用高速 URL，优先路由 — 无速率限制、无认证故障、无延迟。一次性购买，永久使用。

Boost
$2 / 200₽ — one-time

专用高速通道，优先路由。无速率限制，无认证故障，无需等待密钥轮换。

https://boost.gwarden.su/v1
状态 0.00%延迟 -TPS -
获取权限
Dev
$5 / 500₽ — one-time

最高吞吐量，冗余故障转移。为不能容忍延迟的生产级代理设计。

https://dev.gwarden.su/v1
状态 0.00%延迟 -TPS -
获取权限
Emergency
$5 / 500₽ — per month

备用网关：在目录所有模型间切换，主服务整体故障时仍可用。

https://emergency.gwarden.su/v1
状态 100.00%延迟 1.83sTPS -
获取权限
Gwarden AI
$2 / 200₽ — one-time

解锁 Gwarden AI 模型：友好的编程助手，100万上下文。$2 一次性，永久拥有。

"model": "gwarden-ai"
获取权限
价格

订阅

Free
免费
最多 $6$ 信用
每 5 小时 补充至 $6$
7 天试用
Tester
$4
最多 $5$ 信用
每 5 小时 补充至 $5$
1 月$4
2 月$8
3 月$12
4 月$16
5 月$20
6 月$24
7 月$28
8 月$32
9 月$36
10 月$40
11 月$44
1 年$40
1-12 个月。年付按 10 个月计费。
Startupper
$5
最多 $6$ 信用
每 5 小时 补充至 $6$
1 月$5
2 月$10
3 月$15
4 月$20
5 月$25
6 月$30
7 月$35
8 月$40
9 月$45
10 月$50
11 月$55
1 年$50
1-12 个月。年付按 10 个月计费。
Agent
$8
最多 $12$ 信用
每 5 小时 补充至 $12$
1 月$8
2 月$16
3 月$24
4 月$32
5 月$40
6 月$48
7 月$56
8 月$64
9 月$72
10 月$80
11 月$88
1 年$80
1-12 个月。年付按 10 个月计费。
Orchestrator
$15
最多 $20$ 信用
每 5 小时 补充至 $20$
1 月$15
2 月$30
3 月$45
4 月$60
5 月$75
6 月$90
7 月$105
8 月$120
9 月$135
10 月$150
11 月$165
1 年$150
1-12 个月。年付按 10 个月计费。
Neuron
$25
最多 $35$ 信用
每 5 小时 补充至 $35$
1 月$25
2 月$50
3 月$75
4 月$100
5 月$125
6 月$150
7 月$175
8 月$200
9 月$225
10 月$250
11 月$275
1 年$250
1-12 个月。年付按 10 个月计费。
Emperor
$20 $35-43%
最多 $50$ 信用
每 5 小时 补充至 $50$
1 月$20
2 月$40
3 月$60
4 月$80
5 月$100
6 月$120
7 月$140
8 月$160
9 月$180
10 月$200
11 月$220
1 年$200
1-12 个月。年付按 10 个月计费。
General
$55
最多 $65$ 信用
每 5 小时 补充至 $65$
1 月$55
2 月$110
3 月$165
4 月$220
5 月$275
6 月$330
7 月$385
8 月$440
9 月$495
10 月$550
11 月$605
1 年$550
1-12 个月。年付按 10 个月计费。
Developer
$120
无限 - 余额不受限制
无限制
1 月$120
2 月$240
3 月$360
4 月$480
5 月$600
6 月$720
7 月$840
8 月$960
9 月$1080
10 月$1200
11 月$1320
1 年$1200
1-12 个月。年付按 10 个月计费。
CEO
$70
最多 $80$ 信用
每 5 小时 补充至 $80$
1 月$70
2 月$140
3 月$210
4 月$280
5 月$350
6 月$420
7 月$490
8 月$560
9 月$630
10 月$700
11 月$770
1 年$700
1-12 个月。年付按 10 个月计费。
Monopolist
$89
最多 $150$ 信用
每 5 小时 补充至 $150$
1 月$89
2 月$178
3 月$267
4 月$356
5 月$445
6 月$534
7 月$623
8 月$712
9 月$801
10 月$890
11 月$979
1 年$890
1-12 个月。年付按 10 个月计费。

每 5 小时余额补至套餐上限。开始 - 免费 7 天。每 1M token：输入 0.15，输出 0.5，缓存 0.03。充值：1$ = 10.0$。

我无法支付订阅费用，怎么办？

打开 @xRocket 支付指南
快速开始

一分钟启动

1

获取密钥

打开机器人，订阅频道，点击「获取 API」。

2

添加配置

在你的工具中添加 baseURL 和 apiKey。

3

完成

适用于 opencode、Claude Code 和任何 OpenAI 客户端。

一条命令启动 Claude Code

export ANTHROPIC_BASE_URL="https://gwarden.su"
export ANTHROPIC_AUTH_TOKEN="GWAR-XXXX"
export ANTHROPIC_MODEL="glm-5.3"
claude

密钥在个人中心获取。Codex、opencode 和 aider 的配置见文档。

或直接使用 API

# opencode - 检查模型
curl https://gwarden.su/v1/chat/completions \
  -H "Authorization: Bearer GWAR-XXXX" \
  -H "Content-Type: application/json" \
  -d '{"model": "MODEL", "messages": [{"role": "user", "content": "Hello"}]}'