Chat models

Google price change (July 2026): Gemini 2.5 Flash moved from

0.15/

0.60 to $0.30/$ 2.50 per 1M tokens (2–4× increase). Cheaper alternatives: google/gemini-2.5-flash-lite (

0.10/

0.40) or the new google/gemini-3.6-flash.

Frontier (recommended)

ID	Provider	Context	Input $/1M	Output $/1M	Best for
`anthropic/claude-fable-5`	Anthropic	1M	$10	$50	Deep reasoning, long-form writing
`anthropic/claude-opus-5`	Anthropic	1M	$5	$25	Top-tier reasoning, complex code
`anthropic/claude-sonnet-5`	Anthropic	1M	$3	$15	Sweet spot price/quality
`anthropic/claude-haiku-4-5`	Anthropic	200k	$1	$5	Simple tasks, high volume
`google/gemini-3.6-flash`	Google	1M	$1.50	$7.50	Multimodal, long context
`google/gemini-3.5-flash-lite`	Google	1M	$0.30	$2.50	Cheap with 1M context
`google/gemini-3.1-pro-preview`	Google	1M	$2	$12	Preview of new Pro
`google/gemini-2.5-pro`	Google	1M	$1.25	$10	Long context, multimodal
`google/gemini-2.5-flash`	Google	1M	$0.30	$2.50	⚠ price bump Jul 2026
`google/gemini-2.5-flash-lite`	Google	1M	$0.10	$0.40	Cheapest in the catalog
`openai/gpt-5.6-sol`	OpenAI	400k	$5	$30	OpenAI flagship
`openai/gpt-5.6-terra`	OpenAI	400k	$2	$12	Balanced GPT-5.6
`openai/gpt-5.6-luna`	OpenAI	400k	$0.20	$1.20	Cheapest GPT-5.6
`openai/gpt-5.5`	OpenAI	400k	$5	$30	Alternative top-tier
`openai/gpt-5.4-mini`	OpenAI	400k	$0.75	$4.50	Cheap OpenAI
`openai/gpt-5.1`	OpenAI	400k	$1.25	$10	General reasoning
`openai/o3-mini`	OpenAI	200k	$1.10	$4.40	Reasoning (CoT)
`deepseek/deepseek-v4-flash`	DeepSeek	128k	$0.14	$0.28	Open-weight ultra cheap
`deepseek/deepseek-v4-pro`	DeepSeek	128k	$0.44	$0.87	Open-weight reasoning
`moonshot/kimi-k3`	Moonshot	256k	$3	$15	Chinese frontier, code
`moonshot/kimi-k2.7-code`	Moonshot	256k	$0.95	$4	Code-specialized
`moonshot/kimi-k2.7-code-highspeed`	Moonshot	256k	$0.95	$8	Kimi Code low-latency
`xai/grok-4.5`	xAI	500k	$2	$6	Access to X data
`xai/grok-4.3`	xAI	1M	$1.25	$2.50	1M context, cheap
`xai/grok-4.20-reasoning`	xAI	1M	$1.25	$2.50	Cheap reasoning
`xai/grok-build`	xAI	256k	$1	$2	Specialized in dev tasks

Legacy (backward compat)

Still operational for compatibility but not recommended for new integrations:

ID	Suggested replacement
`anthropic/claude-opus-4-8`	`anthropic/claude-opus-5`
`anthropic/claude-sonnet-4-6`	`anthropic/claude-sonnet-5`
`openai/gpt-5`	`openai/gpt-5.6-terra` or `gpt-5.5`
`openai/gpt-4.1-mini`	`openai/gpt-5.4-mini`

Retired in July 2026: openai/gpt-4.1, openai/o4-mini, deepseek/deepseek-chat, deepseek/deepseek-reasoner, moonshot/kimi-k2, moonshot/moonshot-v1-128k, xai/grok-3, xai/grok-3-mini, xai/grok-4.

When to use each

Critical tasks

Claude Fable 5, Claude Opus 5 or GPT-5.6 Sol. Top of class in reasoning.

Production at scale

Claude Sonnet 5. Price/quality balance — pick it blind if you don’t know the rest.

High volume / low cost

Gemini 2.5 Flash Lite, GPT-5.6 Luna or DeepSeek V4 Flash. Sub-dollar per 1M tokens.

Reasoning (chain of thought)

DeepSeek V4 Pro, o3-mini or Grok 4.20 Reasoning. Designed for structured reasoning.

Very long context (1M tokens)

Gemini 3.6 Flash / 2.5 Pro / 2.5 Flash, Grok 4.3 or Claude Fable/Opus/Sonnet 5. Process whole documents.

Code

Kimi K2.7 Code, Claude Sonnet 5 or Grok Build. Strong code performance.

Natural failover

Because all models share the same endpoint and SDK, failover across providers is trivial:

def call_with_fallback(messages):
    for model in ["anthropic/claude-sonnet-5", "openai/gpt-5.6-terra", "google/gemini-2.5-pro"]:
        try:
            return client.chat.completions.create(model=model, messages=messages)
        except Exception:
            continue
    raise RuntimeError("All providers failed")

Get started

Concepts

Advanced features

Legal & compliance

API Reference

Models

Frontier (recommended)

Legacy (backward compat)

When to use each

Critical tasks

Production at scale

High volume / low cost

Reasoning (chain of thought)

Very long context (1M tokens)

Code

Natural failover

​Frontier (recommended)

​Legacy (backward compat)

​When to use each

​Critical tasks

​Production at scale

​High volume / low cost

​Reasoning (chain of thought)

​Very long context (1M tokens)

​Code

​Natural failover

Frontier (recommended)

Legacy (backward compat)

When to use each

Critical tasks

Production at scale

High volume / low cost

Reasoning (chain of thought)

Very long context (1M tokens)

Code

Natural failover