Model catalogue
Every major model on one key
Claude for reasoning, DeepSeek for code, Qwen for value, Kimi and GLM for multilingual workflows. Claude & GPT billed at ; other major models at competitive discounts. Live list per key via GET /v1/models.
Claude Opus 5
Anthropic · Flagship
Per 1M tokens
Live
Claude Opus 5.5
Anthropic · Flagship
Per 1M tokens
Live
Claude Opus 4.8
Anthropic · Flagship
Per 1M tokens
Live
Claude Opus 4.7
Anthropic · Flagship
Per 1M tokens
Live
Claude Opus 4.6
Anthropic · Flagship
Per 1M tokens
Live
Claude Fable 5
Anthropic · Flagship
Per 1M tokens
Live
Claude Sonnet 5
Anthropic · Balanced
Per 1M tokens
Live
Claude Sonnet 4.6
Anthropic · Balanced
Per 1M tokens
Live
Claude Haiku 4.5
Anthropic · Fast & light
Per 1M tokens
Live
GPT-5.5
OpenAI · Flagship
Per 1M tokens
Live
GPT-5.6 Sol
OpenAI · Flagship
Per 1M tokens
Live
GPT-5.6 Terra
OpenAI · Balanced
Per 1M tokens
Live
GPT-6 Astra
OpenAI · Flagship
Per 1M tokens
Live
GPT-6 Sol
OpenAI · Flagship
Per 1M tokens
Live
GPT-6.1 Sol
OpenAI · Flagship
Per 1M tokens
Live
GPT-5.3 Codex Spark
OpenAI · Coding
Per 1M tokens
Live
G
Gemini 3.5 Flash
Google · Fast & light
Per 1M tokens
Live
G
Gemini 3.1 Pro (prev.)
Google · Flagship
Per 1M tokens
Live
G
Gemini 3 Flash (prev.)
Google · Fast & light
Per 1M tokens
Live
G
Gemini 2.5 Pro
Google · Flagship
Per 1M tokens
Live
G
Gemini 2.5 Flash
Google · Fast & light
Per 1M tokens
Live
G
Gemini 2.5 Flash Lite
Google · Fast & light
Per 1M tokens
Live
Gemini 3 Pro Image
Google · Image
Per image
Live
Gemini 3.1 Flash Image
Google · Image
Per image
Live
Gemini 3.1 Flash Lite Image
Google · Fast & light
Per image
Live
Gemini Nano Banana 2.1
Google · Image
Per image
Live
DeepSeek V4.1 Flash
DeepSeek · Fast & light
Per 1M tokens
Live
DeepSeek V4 Pro 0813
DeepSeek · Flagship
Per 1M tokens
Live
DeepSeek V4 Pro
DeepSeek · Flagship
Per 1M tokens
Live
Qwen3.8 Max
Alibaba Qwen · Flagship
Per 1M tokens
Live
Qwen3.8 Max 0902
Alibaba Qwen · Flagship
Per 1M tokens
Live
Qwen3.8 Flash
Alibaba Qwen · Fast & light
Per 1M tokens
Live
Qwen3.7 Max
Alibaba Qwen · Flagship
Per 1M tokens
Live
Qwen3.7 Max 06-08
Alibaba Qwen · Flagship
Per 1M tokens
Live
Qwen3.7 Plus
Alibaba Qwen · Balanced
Per 1M tokens
Live
Qwen3.7 Plus 05-26
Alibaba Qwen · Balanced
Per 1M tokens
Live
Qwen3.7 Flash
Alibaba Qwen · Fast & light
Per 1M tokens
Live
Qwen3.6 Plus
Alibaba Qwen · Balanced
Per 1M tokens
Live
Qwen3-VL Plus
Alibaba Qwen · Vision
Per 1M tokens
Live
Kimi K3
Moonshot · Long context
Per 1M tokens
Live
Kimi K2.7 Code
Moonshot · Coding
Per 1M tokens
Live
GLM-5.3
Zhipu GLM · Balanced
Per 1M tokens
Live
GLM-5.3 Flash
Zhipu GLM · Fast & light
Per 1M tokens
Live
M
MiniMax M3
MiniMax · Balanced
Per 1M tokens
Live
Grok Imagine Image
xAI · Image
Per image
Live
Grok Imagine Image Quality
xAI · Image
Per image
Live
Grok Imagine Video
xAI · Video
Per 5s video
Live
How usage is billed
Usage-based, metered per call. No subscription, credits never expire.
Input tokens 1×
Prompt and context sent to the model. Billed at each model's published rate — Claude & GPT at .
Output tokens 5× input
Generated text. Most models bill output at 5× the input rate (DeepSeek 2×). Streaming replies count the same as non-streaming.
Cache hit (read) provider's cache rate
Reused context from the provider's prompt cache — charged at the cache-read rate each model provider publishes. For example, Claude series cache reads are 10% of the input rate, while DeepSeek, GLM, MiniMax and others each follow their own official cache pricing. Keep your system prompt stable to hit the cache; this is the single biggest cost saver.
Cache write Claude: 1.25× input
Storing context for reuse. For the Claude series, cache writes are charged once at 125% of the input rate (and Claude cache reads are then 10%). Other providers follow their own official cache-write pricing.
Thinking tokens = output
Reasoning tokens from extended-thinking models are billed exactly like output tokens — no hidden multiplier.
Discounts standing
Discounts are policy, not promotions: Claude & GPT at ; DeepSeek · Kimi at ; GLM from 15% off; Qwen at . They do not expire and are not tied to a promotion window.
Every call is itemised in the console — input, output, cache and thinking tokens per request. Verify live rates with GET /v1/models.