Canonical model records

Models

The scoring unit is a named release, or an explicit internal/unreleased system when a lab has not named the model. See also families.

• GPT-6 Astra

  OpenAI · GPT-6
  20.60

  +20.60
• OpenAI Internal Research System — 2026-09

  OpenAI · GPT-6

  Internal / unreleased
  18.56

  +18.56
• ChatGPT

  OpenAI · GPT
  17.26

  +18.93
• AlphaFold 3

  Google DeepMind · AlphaFold
  10.77

  +10.77
• RFdiffusion

  University of Washington Institute for Protein Design · RoseTTAFold
  8.45

  +8.45
• OpenAI Internal Research System — 2026-05

  OpenAI · GPT-6

  Internal / unreleased
  6.12

  +6.12
• Anthropic Internal Research Model — 2026-08

  Anthropic · Claude

  Internal / unreleased
  5.06

  +5.06
• Gemini 3.1 Pro

  Google DeepMind · Gemini
  4.13

  +4.13
• OpenAI Internal Research System — October 2026 Scientific Results

  OpenAI · OpenAI Internal Research

  Internal / unreleased
  3.98

  +3.98
• AIFS

  European Centre for Medium-Range Weather Forecasts · AIFS
  3.68

  +3.68
• PaLM 2

  Google DeepMind · PaLM
  3.54

  +3.54
• AlphaGenome

  Google DeepMind · AlphaGenome
  3.27

  +3.27
• AlphaMissense

  Google DeepMind · AlphaFold
  2.65

  +2.65
• GraphCast

  Google DeepMind · GraphCast
  2.65

  +2.65
• OpenAI Codex

  OpenAI · Codex
  2.47

  +2.49
• GenCast

  Google DeepMind · GraphCast
  2.46

  +2.46
• WeatherNext Cyclones

  Google DeepMind · GraphCast
  2.43

  +2.43
• GNoME

  Google DeepMind · GNoME
  2.38

  +2.38
• AlphaChip

  Google DeepMind · AlphaChip
  2.31

  +2.31
• Waymo Driver

  Waymo · Waymo Driver
  2.16

  +2.16
• ESM3

  EvolutionaryScale · ESM
  2.10

  +2.10
• Google Flood Forecasting Model

  Google · NeuralHydrology
  2.00

  +2.00
• RoseTTAFold All-Atom

  University of Washington Institute for Protein Design · RoseTTAFold
  1.94

  +1.94
• MatterGen

  Microsoft · MatterGen
  1.74

  +1.74
• Pangu-Weather

  Huawei · Pangu
  1.73

  +1.73
• Evo

  Arc Institute · Evo
  1.67

  +1.67
• Profluent Protein Language Model

  Profluent · Profluent PLM
  1.67

  +1.67
• Aurora

  Microsoft · Aurora
  1.64

  +1.64
• NeuralGCM

  Google DeepMind · NeuralGCM
  1.56

  +1.56
• Claude 3.7 Sonnet

  Anthropic · Claude
  1.37

  +1.39
• AlphaProteo

  Google DeepMind · AlphaProteo
  1.22

  +1.22
• Anthropic Claude Science Internal Research Model — Scorpion Binder Campaign (2026)

  Anthropic · Claude

  Internal / unreleased
  1.22

  +1.22
• codex-1

  OpenAI · Codex
  1.16

  +1.16
• Gemini

  Google DeepMind · Gemini
  1.13

  +1.13
• Claude Fable 5

  Anthropic · Claude
  1.12

  +1.12
• GPT-5 Pro

  OpenAI · GPT
  1.01

  +1.01
• AlphaDev

  Google DeepMind · AlphaZero
  0.96

  +0.96
• BindCraft

  EPFL · BindCraft
  0.89

  +0.89
• GPT-5.6

  OpenAI · GPT-5.6

  Version unspecified
  0.66

  +0.66
• Queen of Hearts

  Powerful Medical · Queen of Hearts
  0.57

  +0.57
• Boltz-1

  Massachusetts Institute of Technology · Boltz
  0.52

  +0.52
• AlphaFold 2

  Google DeepMind · AlphaFold
  0.51

  +0.51
• GPT-5.6 Pro

  OpenAI · GPT-5.6
  0.48

  +0.48
• WeatherNext 3

  Google DeepMind · GraphCast
  0.42

  +0.42
• GPT-5.6 Sol

  OpenAI · GPT-5.6
  0.40

  +1.48
• GPT-4

  OpenAI · GPT
  0.40

  +0.40
• Literature NLP Recipe Model

  Lawrence Berkeley National Laboratory · Materials NLP
  0.37

  +0.37
• Chemprop Ensemble Internal Research Model

  Massachusetts Institute of Technology · Chemprop

  Internal / unreleased
  0.35

  +0.35
• Claude

  Anthropic · Claude

  Version unspecified
  0.32

  +3.30
• Chai-1

  Chai Discovery · Chai
  0.32

  +0.32
• GPT-5

  OpenAI · GPT
  0.32

  +0.32
• AMIE

  Google DeepMind · Gemini/PaLM medical
  0.29

  +0.29
• Gemini Deep Think

  Google DeepMind · Gemini
  0.27

  +0.27
• Evo 2

  Arc Institute · Evo
  0.24

  +0.24
• WeatherNext 2

  Google DeepMind · GraphCast
  0.24

  +0.24
• SL2T

  Google DeepMind · SL2T
  0.24

  +0.24
• Pharma.AI Internal Research Model

  Insilico Medicine · Pharma.AI

  Internal / unreleased
  0.17

  +0.17
• AlphaEarth Foundations

  Google DeepMind · AlphaEarth
  0.16

  +0.16
• AlphaProof

  Google DeepMind · AlphaProof
  0.14

  +0.14
• Ganeshram–Duruisseaux–Anandkumar PINN

  California Institute of Technology · PINN
  0.12

  +0.12
• Chemprop-style Internal Research Model

  Massachusetts Institute of Technology · Chemprop

  Internal / unreleased
  0.08

  +0.08
• Green Light Traffic Model

  Google · Google Maps models
  0.07

  +0.07
• Gemini 2.0

  Google DeepMind · Gemini
  0.04

  +0.04
• Gemini 3.0 Flash

  Google DeepMind · Gemini
  0.03

  +0.03
• Air Canada Website Chatbot

  Air Canada · Unknown commercial chatbot
  -0.11

  0.00
• Tessa

  National Eating Disorders Association · Cass chatbot
  -0.12

  0.00
• ElevenLabs Voice Model

  ElevenLabs · ElevenLabs
  -0.18

  0.00
• Anthropic Internal Research Test Model — 2026

  Anthropic · Claude

  Internal / unreleased
  -0.23

  0.00
• Claude Opus 4.6

  Anthropic · Claude
  -0.23

  0.00
• Grok

  xAI · Grok
  -0.24

  0.00
• Microsoft Designer

  Microsoft · Designer
  -0.26

  0.00
• Claude Opus 4.7

  Anthropic · Claude
  -0.48

  +0.08
• Claude Mythos 5

  Anthropic · Claude
  -0.76

  0.00
• OpenAI Internal Research System — 2026-05 (wiki swarm)

  OpenAI · OpenAI Internal

  Internal / unreleased
  -0.78

  0.00
• Unknown Deepfake System

  Unknown · Unknown generative media models
  -0.84

  0.00
• nH Predict

  UnitedHealth Group · nH Predict
  -1.26

  0.00
• Character.AI Dialogue Model

  Character.AI · Character.AI
  -1.54

  0.00
• OpenAI Internal Research System — 2026-07 (cyber evals)

  OpenAI · OpenAI Internal

  Internal / unreleased
  -1.66

  0.00
• Unspecified AI system

  Unknown · Unspecified

  Version unspecified
  -2.32

  +2.86
• GPT-4o

  OpenAI · GPT
  -2.76

  +0.01
• Cruise ADS

  Cruise · Cruise autonomous driving system
  -3.08

  0.00