Models

AllToken aggregates 300+ models from 60+ providers. Each model has different strengths in cost, speed, context window, and capabilities.

Browse the full catalog at Models or query programmatically via the Models API.

Each model has a unique ID used in the model parameter of your requests:

Common model IDs include:

• gpt-4o — OpenAI GPT-4o (multimodal, fast)
• claude-sonnet-4 — Anthropic Claude Sonnet 4 (reasoning, coding)
• deepseek-chat — DeepSeek Chat (cost-effective)
• deepseek-reasoner — DeepSeek Reasoner (extended thinking)
• gemini-2.5-flash — Google Gemini 2.5 Flash (fast, multimodal)

Models vary in their supported capabilities:

• Text — standard text generation (all models)
• Multimodal — accepts image inputs alongside text
• Reasoning — extended thinking / chain-of-thought
• Function Calling — structured tool use
• Web Search — real-time internet access
• JSON Mode — guaranteed JSON output
• Streaming — token-by-token SSE responses

Filter by capability on the Models page or check a model's detail page for the full list.

Pricing is per million tokens, billed separately for input and output. Prices vary by model and provider:

• Budget models (DeepSeek Chat) — from $0.07 / $0.28 per 1M tokens
• Mid-range models (Claude Haiku, GPT-4o mini) — $0.25-$1.00 per 1M tokens
• Flagship models (Claude Opus, GPT-4o) — $2.50-$15.00 per 1M tokens
• Free tier — some models available at no cost with rate limits

See live pricing on each model's detail page, or use Cost Tracking to monitor spend.

The context window is the maximum number of tokens (input + output) a model can process in a single request. Common sizes:

• 128K tokens — GPT-4o, Claude Sonnet 4, DeepSeek Chat
• 200K tokens — Claude Haiku 4.5, Claude Opus 4
• 1M tokens — Claude Opus 4.6, Gemini 2.5 Pro

Requests that exceed the context window return an error. Summarize long documents or break them into chunks to stay within limits.