Skip to content

Supported Models

Axiom aggregates models from Anthropic, OpenAI, Google, xAI, and major Chinese providers, all callable through a unified OpenAI-compatible interface. The Model Name in the tables below is the value to put in the model field of your request.

Live models and pricing

The model list is continuously updated. For the most current full list and real-time ratios, refer to the "Model Pricing" page in the console: axiomcode.top.

Anthropic · Claude

Model NameDescription
claude-fable-5(Temporarily unavailable due to US government policy) Fable series
claude-opus-4-8Opus flagship; top choice for complex reasoning and long-form writing
claude-opus-4-7Previous-generation Opus
claude-opus-4-6Early Opus version
claude-sonnet-4-6Balanced performance and cost; suited for everyday use and coding
claude-haiku-4-5-20251001Lightweight and ultra-fast; suited for high concurrency and short tasks

OpenAI · GPT

Model NameTypeDescription
gpt-5.5ChatFlagship general-purpose model
gpt-5.5-openai-compactChatCompact version
gpt-5.4ChatPrevious-generation flagship
gpt-5.4-miniChatLightweight and low-cost
gpt-5.4-openai-compactChatCompact version
codex-auto-reviewCodingCode review
gpt-image-2ImageText-to-image
codex-gpt-image-2ImageImage generation for coding scenarios

Google · Gemini

Model NameTypeDescription
gemini-3.1-pro-preview-thinkingReasoningLatest Pro thinking model
gemini-3-pro-preview-thinkingReasoningPro thinking model (includes -128 / -512 thinking-budget variants)
gemini-3.1-pro-previewChatLatest Pro
gemini-3-pro-previewChatPro
gemini-3-flash-previewChatFlash; fast responses
gemini-3.5-flash-cChatFlash with high caching (-low-c is the low-cache variant)
gemini-2.5-proChatPrevious-generation Pro (-thinking is the thinking version)
gemini-2.5-flashChatPrevious-generation Flash
gemini-2.5-flash-liteChatMore lightweight Flash
gemini-3-pro-image-previewImageNative multimodal image generation

xAI · Grok

Model NameTypeDescription
grok-3ChatGeneral-purpose chat
grok-420-fastChatFast version
grok-420-fast-deepsearchWeb searchBuilt-in deep web search
grok-imagine-videoVideoText-to-video
grok-imagine1.5-videoVideoText-to-video 1.5
grok-imagine-video-1.5-previewVideoText-to-video 1.5 preview

Chinese Models

Model NameProviderDescription
deepseek-v4-proDeepSeekFlagship reasoning
deepseek-v4-flashDeepSeekFast version
qwen3.7-maxTongyi QianwenFlagship
qwen3.7-plus / qwen3.6-plus / qwen3.5-plusTongyi QianwenPlus series
glm-5.1 / glm-5Zhipu GLMGeneral-purpose chat
kimi-k2.6 / kimi-k2.5Moonshot KimiLong context
minimax-m3 / minimax-m2.7 / minimax-m2.5MiniMaxGeneral-purpose chat
hy3-previewTencent HunyuanPreview version

Xiaomi · MiMo

Model NameTypeDescription
mimo-v2.5-pro / mimo-v2-proChatPro chat
mimo-v2.5ChatStandard chat
mimo-v2-omniMultimodalOmni-modal understanding
mimo-v2.5-asrSpeech recognitionSpeech-to-text
mimo-v2.5-tts / mimo-v2-ttsSpeech synthesisText-to-speech
mimo-v2.5-tts-voicecloneSpeech synthesisVoice cloning
mimo-v2.5-tts-voicedesignSpeech synthesisVoice design

Model groups (group ratios)

On the backend, models are divided into different groups by channel (such as Claude - Max, 哈吉米 Gemini, 国产模型 (Chinese Models), etc.). Different groups correspond to different stability and price ratios, which you can select when creating a key in the console. For billing details, see Recharge & Subscription.