Every model on the bench, measured.

Anthropic's most capable model — the reliability frontier for the hardest agentic work.

State-of-the-art autonomous coding reliability at half the price of Fable 5.

Previous-generation frontier Opus, still highly reliable for agentic coding.

The production workhorse — best balance of speed and intelligence.

The most agentic Sonnet yet — near-Opus coding reliability at Sonnet cost.

Fastest, cheapest Claude — snappy everyday tasks at scale.

OpenAI's GPT-5.6 flagship — its best coding-and-agentic model yet, for the hardest work.

The GPT-5.6 balanced workhorse — GPT-5.5-class quality at half the price.

The fast, cost-efficient GPT-5.6 tier for everyday work at scale.

Sol's maximum-compute mode — cooperating sub-agents for the hardest multi-step work.

Previous-generation OpenAI flagship — frontier-reliable, now succeeded by GPT-5.6 Sol.

Previous-generation maximum-accuracy OpenAI tier, succeeded by Sol's Ultra mode.

Previous-generation balanced workhorse, succeeded by GPT-5.6 Terra.

Low-cost previous-generation OpenAI workhorse for standard routing.

OpenAI's cheapest small model for high-volume simple work.

Google's newest flagship — a 2M-token context and a Deep Think reasoning mode.

Top-tier reasoning, knowledge and vision with a 1M-token multimodal context.

Fast, cheap multimodal workhorse with near-Pro quality.

Cheapest Gemini, optimized for low-latency high-volume traffic.

xAI's flagship: strong math reasoning and real-time knowledge at aggressive pricing.

Cheap, fast, 2M-context workhorse for high throughput.

Open natively-multimodal MoE flagship — a strong, cheap general workhorse.

Ultra-long 10M-token context, efficient single-GPU open model.

Europe's flagship open-weight MoE — a strong, cheap, multilingual generalist.

Efficient dense workhorse with a built-in coding agent.

Tiny Apache-2.0 multimodal model unifying reasoning, vision and coding.

Mistral's transparent-chain reasoning flagship, now multimodal.

Open MIT flagship — 1.6T-parameter MoE with native vision and a 1M-token context at throwaway prices.

Cheap, fast open MoE — near-frontier value at a fraction of Pro's cost and size.

Ultra-cheap open MoE with standout math and efficiency.

Iconic open reasoning model — strong math at low cost, now a generation behind.

Alibaba's proprietary agentic reasoning flagship with a 1M context.

Open-weight reasoning model with strong math and science.

Small, fast open-weight MoE with big-model manners.

Best open-weights Qwen for agentic coding — great value.

Open-weight coding flagship at a fraction of frontier cost.

Efficient open coding workhorse for everyday dev tasks.

Open-weight coding refresh of Kimi K2 — stronger agentic tool use at low cost.

Open-weight agentic powerhouse with long context and video input.

Deep-reasoning open variant of Kimi K2 for hard problems.

Open-weight multimodal model with native video input and 1M context.