51 models · 7 benchmarks · 13 providers

Every AI model. One honest scoreboard.

Benchmarks, API pricing, context windows and capabilities for GPT, Claude, Gemini, DeepSeek, Llama and everything in between — side-by-side, kept current, free forever. No account, no paywall.

Browse the catalog →Estimate your API costs →Free JSON API →

Frontier right now

Top models by modhub Index — the average across all tracked benchmarks.

All leaderboards →

1OpenAI o4-miniOpenAI$1.1/$4.4 · 200K80.9 2Qwen3-MaxAlibaba (Qwen)$1.2/$6 · 262K80.5 3GPT-5.5OpenAI$5/$30 · 1M77.2 4Gemini 3 ProGoogle$2/$12 · 1M73.6 5Gemini 2.5 ProGoogle$1.25/$10 · 1.0M70.9 6GPT-5OpenAI$1.25/$10 · 400K70.7 7Qwen3-235B-A22BAlibaba (Qwen)$0.22/$0.88 · 262K69.0 8OpenAI o3OpenAI$2/$8 · 200K68.9 9MiniMax M2MiniMax$0.3/$1.2 · 205K68.9 10DeepSeek V3.2DeepSeek$0.28/$0.42 · 128K68.4

Head-to-head comparisons

Specs, benchmarks and pricing — settled side by side.

The flagship fight

Claude Opus 4.8

GPT-5.5

See the breakdown →

The workhorse duel

Claude Sonnet 4.6

GPT-5.2

See the breakdown →

Speed vs muscle

Gemini 3.5 Flash

GPT-5.5

See the breakdown →

Closed vs open frontier

Claude Opus 4.8

DeepSeek V4

See the breakdown →

Open-weights heavyweights

DeepSeek V4

Kimi K2 Thinking

See the breakdown →

Budget tier battle

Claude Haiku 4.5

Gemini 2.5 Flash

See the breakdown →

Cost calculator

Model your real workload — tokens, volume, cache hit rate — and rank every API by monthly cost. Presets for chatbots, RAG and coding agents.

Coding agent catalog

Claude Code, Cursor, Codex, Devin and a dozen more — categorized by how they work, what they cost and which models they can drive.

Free JSON API

The entire dataset — models, pricing, benchmark scores — as clean JSON. No key, no rate-limit drama. Build your own dashboards on it.