Swap between Fable, Claude, GPT, Gemini, DeepSeek, GLM and Grok by changing a single string. Pricing, context limits, modalities and streaming support come from the catalog.
AnthropicFable
Long-horizon reasoning and agentic tool use. Holds context across multi-step work without losing the thread.
POST /v1/chat/completionsAnthropicClaude
Frontier depth on hard problems — research, refactors and analysis where a wrong answer costs more than a slow one.
POST /v1/chat/completions
OpenAIGPT
Broad general capability with strong instruction following and native structured output.
POST /v1/chat/completions
GoogleGemini
A two-million-token window with native multimodal input. The choice when the whole corpus has to fit in one prompt.
POST /v1/chat/completionsAnthropicClaude
The default for production traffic. Near-flagship quality at a fifth of the cost, with a million-token window.
POST /v1/chat/completions
DeepSeekDeepSeek
Open-weight reasoning at a fraction of frontier pricing. Strong on mathematics and competitive programming.
POST /v1/chat/completions
xAIGrok
Fast reasoning with a wide context window and a distinctly direct answering style.
POST /v1/chat/completions
Z.aiGLM
Open-weight and fast. The cost floor for classification, extraction and high-volume batch work.
POST /v1/chat/completionsZ.aiGLM
Text embeddings at 1024 dimensions. Batch up to 128 inputs per request for retrieval, clustering and classification.
POST /v1/embeddings