MMLU-Pro benchmark
267 AI models ranked on MMLU-Pro. Each row shows the model’s score next to its cheapest API price per 1M tokens where it has one, so you can weigh quality against cost.
Scores frozen
Gemini 3 Pro (Google) leads at 90%
Models ranked
267
on this benchmark
Top score
90%
on the frozen board
Type
Single eval
benchmark
Frozen
Aug 18, 2026
benchmark scores
Cost vs MMLU-Pro
171 priced modelsEach labelled name links to that model’s page. Curated to the cost-efficiency frontier, top scorers and cheapest so labels stay readable; the full field is in the leaderboard table below (and via “select all”). Reasoning models (extended “thinking” before answering) carry a ring around the dot.
Models ranked on MMLU-Pro
highest score first · 267 models| # | ||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Gemini 3 Pro |  Google | 90% | $2.00 via Replicate | $12.00 via Replicate | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 2 | Claude Opus 4.5 |  Anthropic | 89%reasoning: true | $5.00 via Anthropic | $25.00 via Anthropic | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 3 | Gemini 3 Flash | Google | 89%reasoning: true | $0.50 Routed quote | $3.00 Routed quote | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 4 | Claude Opus 4.1 | Anthropic | 88%reasoning: true | $15.00 via Anthropic | $75.00 via Anthropic | |||||||||||||||||||||
| 5 | MiniMax M2.1 |  MiniMax | 87% | $0.30 via Minimax | $1.20 via Minimax | |||||||||||||||||||||
| 6 | Claude 4.5 Sonnet | Anthropic | 87%reasoning: true | $3.00 via Replicate | $15.00 via Replicate | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 7 | GPT-5.2 |  OpenAI | 87% | $1.75 via OpenAI | $14.00 via OpenAI | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 8 | Claude 4 Opus | Anthropic | 87%reasoning: true | $5.00 via Snowflake | $25.00 via Snowflake | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 9 | GPT-5 | OpenAI | 87% | $1.25 via Databricks | $10.00 via Databricks | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 10 | GPT-5.1 | OpenAI | 87% | $1.25 via Databricks | $10.00 via Databricks | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 11 | Grok 4 |  xAI | 87% | $1.25 via xAI | $2.50 via xAI | |||||||||||||||||||||
| 12 | GPT 5 Codex | OpenAI | 86% | $1.25 Routed quote | $10.00 Routed quote | |||||||||||||||||||||
| 13 | DeepSeek V3.2 Speciale |  DeepSeek | 86% | $0.58 via Azure | $1.68 via Azure | |||||||||||||||||||||
| 14 | DeepSeek V3.2 | DeepSeek | 86%reasoning: true | $0.259 Routed quote | $0.38 via DeepInfra | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 15 | Gemini 2.5 Pro | Google | 86% | $1.25 via Databricks | $10.00 via Databricks | |||||||||||||||||||||
| 16 | GPT-5.1-Codex | OpenAI | 86% | $1.25 Routed quote | $10.00 Routed quote | |||||||||||||||||||||
| 17 | Gemini 2.5 Pro (Mar) | Google | 86% | n/a | n/a | |||||||||||||||||||||
| 18 | GLM 4.7 |  Zhipu | 86% | $0.40 via DeepInfra | $1.75 via DeepInfra | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 19 | Doubao Seed Code | ByteDance Seed | 85% | n/a | n/a | |||||||||||||||||||||
| 20 | Grok 4.1 Fast Reasoning | xAI | 85% | $0.20 via Azure | $0.50 via Azure | |||||||||||||||||||||
| 21 | o3 | OpenAI | 85% | $2.00 via OpenAI | $8.00 via OpenAI | |||||||||||||||||||||
| 22 | DeepSeek V3.1 Terminus | DeepSeek | 85%reasoning: true | $0.27 via DeepInfra | $1.00 via DeepInfra | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 23 | DeepSeek V3.1 | DeepSeek | 85%reasoning: true | $0.25 via DeepInfra | $0.95 via DeepInfra | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 24 | Grok 4 Fast Reasoning | xAI | 85% | $0.20 via Azure | $0.50 via Azure | |||||||||||||||||||||
| 25 | DeepSeek V3.2 Exp (0925) (Reasoning) | DeepSeek | 85% | n/a | n/a | |||||||||||||||||||||
| 26 | Cogito 671B V2 P1 | Other | 85%reasoning: true | $1.20 via Fireworks AI | $1.20 via Fireworks AI | |||||||||||||||||||||
| 27 | Kimi K2 Thinking |  Moonshot | 85% | $0.60 via Moonshot | $1.20 via Gmi | |||||||||||||||||||||
| 28 | O1 Preview | OpenAI | 85% | $16.50 via Azure | $66.00 via Azure | |||||||||||||||||||||
| 29 | Mimo V2 Flash |  Xiaomi | 84%reasoning: true | $0.10 Routed quote | $0.30 Routed quote | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 30 | Qwen3 235B A22b Instruct 2507 |  Alibaba | 84%reasoning: true | $0.09 via DeepInfra | $0.10 via Wandb | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 31 | Gemini 2.5 Flash Preview 09 2025 | Google | 84%reasoning: true | $0.30 via Google | $2.50 via Google | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 32 | Claude 4 Sonnet | Anthropic | 84%reasoning: true | $3.00 via Replicate | $15.00 via Replicate | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 33 | Qwen3 Max | Alibaba | 84% | $0.78 Routed quote | $3.90 Routed quote | |||||||||||||||||||||
| 34 | o1 | OpenAI | 84% | $15.00 via OpenAI | $60.00 via OpenAI | |||||||||||||||||||||
| 35 | Qwen3 Max (Preview) | Alibaba | 84% | $1.20 Routed quote | $6.00 Routed quote | |||||||||||||||||||||
| 36 | K-EXAONE | Lg-ai-research | 84% | n/a | n/a | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 37 | GPT-5 Mini | OpenAI | 84% | $0.25 via Databricks | $1.20 via Snowflake | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 38 | Gemini 2.5 Pro (May) | Google | 84% | n/a | n/a | |||||||||||||||||||||
| 39 | Claude 3.7 Sonnet | Anthropic | 84%reasoning: true | $3.00 via Databricks | $15.00 via Replicate | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 40 | DeepSeek V3.2 Exp (0925) | DeepSeek | 84% | n/a | n/a | |||||||||||||||||||||
| 41 | Qwen3 VL 235B A22b | Alibaba | 84%reasoning: true | $0.53 via Amazon Bedrock | $2.66 via Amazon Bedrock | |||||||||||||||||||||
| 42 | GLM 4.5 | Zhipu | 84% | $0.40 via DeepInfra | $1.60 via DeepInfra | |||||||||||||||||||||
| 43 | o4 Mini | OpenAI | 83% | $1.00 via Replicate | $4.00 via Replicate | |||||||||||||||||||||
| 44 | Gemini 2.5 Flash | Google | 83%reasoning: true | $0.15 via OCI | $0.60 via OCI | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 45 | ERNIE 5.0 Thinking Preview |  Baidu | 83% | n/a | n/a | |||||||||||||||||||||
| 46 | Nova 2.0 Pro Preview |  Amazon | 83%medium | n/a | n/a | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 47 | GLM 4.6 | Zhipu | 83%reasoning: true | $0.50 via DeepInfra | $2.00 via DeepInfra | |||||||||||||||||||||
| ||||||||||||||||||||||||||
| 48 | Grok 3 Mini | xAI | 83%reasoning: true | $0.25 via Azure | $0.50 via OCI | |||||||||||||||||||||
| 49 | Llama 3.1 Nemotron Ultra 253B V1 |  NVIDIA | 82%reasoning: true | $0.60 via Nebius | $1.80 via Nebius | |||||||||||||||||||||
| 50 | Qwen3 Max Thinking (Preview) | Alibaba | 82% | n/a | n/a | |||||||||||||||||||||
Frequently asked questions
What is the MMLU-Pro benchmark?
A harder, cleaned-up successor to MMLU testing multi-task knowledge and reasoning across academic subjects. It expands each question from four to ten answer options and removes trivial or noisy items, making it more reasoning-intensive and prompt-stable. Scored as multiple-choice accuracy.
Which AI model scores highest on MMLU-Pro?
Gemini 3 Pro (Google) leads with 90%, from $2.00 per 1M input tokens from Replicate. Artificial Analysis last published scores for this benchmark on Aug 18, 2026. Scores here are frozen at that date, and models released since are not on it.
How many models are ranked on MMLU-Pro?
267 models carry an MMLU-Pro score. Scores come from Artificial Analysis; where a model has a listed API price it is drawn from LiteLLM and OpenRouter, and models with no current price are listed without one.
Other benchmarks
compare the same models on a different evalEvery weekday
AI moves fast. Here's your debrief.
News, analysis, tools, and more.
For people who build with AI