Pass IndexThe State of AISign in

Weirdml — Epoch AI

Weirdml — Epoch AI is run by Epoch AI. It has ranked 170 models, of which the catalogue holds 117, scoring from 1.73 to 92.9.

measured by Epoch AI · the board itself

PlaceModelMetricScore
1stof 170Claude Fable 5.1Accuracy (max)92.9
2ndof 170GPT 6 AstraAccuracy (high)92.87
4thof 170Claude Fable 5Accuracy (max)91.94
5thof 170Claude Opus 5Accuracy (max)91.78
8thof 170GPT-5.6 SolAccuracy (high)88.76
12thof 170GPT-5.5Accuracy (xhigh)84.91
14thof 170Claude Opus 4.8Accuracy (xhigh)82.89
15thof 170Kimi K3Accuracy (max)82.57
16thof 170GPT-5.3 CodexAccuracy79.3
17thof 170GPT-5.6 TerraAccuracy (high)78.27
18thof 170Claude Opus 4.6Accuracy (high)77.95
21stof 170GPT-5.4Accuracy (xhigh)77.7
22ndof 170Claude Opus 4.7Accuracy (high)76.44
27thof 170GLM-5.3Accuracy (max)75.4
28thof 170GPT-5.2Accuracy (xhigh)72.19
29thof 170Gemini 3.1 Pro PreviewAccuracy72.07
31stof 170GLM 5.2Accuracy (max)70.12
32ndof 170gemini-3-pro-previewAccuracy69.93
33rdof 170Claude Sonnet 5Accuracy (high)68.78
35thof 170Grok 4.6Accuracy (high)67.29
37thof 170DeepSeek V4 Pro 0813Accuracy (max)66.22
38thof 170Claude Sonnet 4.6Accuracy (medium)66.07
40thof 170Claude Opus 4.5Accuracy63.74
42ndof 170DeepSeek V4 Flash (0731)Accuracy (max)62.96
43rdof 170Gemini 3.5 FlashAccuracy (high)62.64
44thof 170Gemini 3 Flash PreviewAccuracy61.6
45thof 170GPT-5.6 LunaAccuracy (high)60.86
46thof 170GPT-5.1Accuracy (high)60.77
47thof 170GPT-5Accuracy (high)60.7
48thof 170GPT-5 ProAccuracy (high)60.39
49thof 170GPT-5.4 miniAccuracy (high)60.3
50thof 170Muse Spark 1.2Accuracy (xhigh)60.3
51stof 170o3-proAccuracy (high)58.21
53rdof 170GPT-5.4 ProAccuracy (none)57.44
54thof 170GLM 5.1Accuracy57.1
56thof 170Gemini 3.6 FlashAccuracy (high)56.1
57thof 170Kimi K2.6Accuracy55.86
58thof 170GPT-5 Codex (batch)Accuracy54.53
60thof 170Kimi K2.7 CodeAccuracy54.12
61stof 170Gemini 2.5 ProAccuracy54.03
62ndof 170GPT-5 miniAccuracy (high)52.67
63rdof 170o4-miniAccuracy (high)52.56
64thof 170o3Accuracy (high)52.42
65thof 170Grok 4.20Accuracy52.26
66thof 170Gemma 4 31BAccuracy52.26
67thof 170Gemini 3.1 Flash-LiteAccuracy52.19
68thof 170Grok 4.3Accuracy49.89
71stof 170GPT-5.4 nanoAccuracy (high)49.23
72ndof 170DeepSeek V4 ProAccuracy (max)48.9
73rdof 170GPT OSS 120BAccuracy (high)48.17
75thof 170GLM-5Accuracy48.17
76thof 170Claude Sonnet 4.5Accuracy47.71
77thof 170o1Accuracy47.56
78thof 170DeepSeek-V3.2-SpecialeAccuracy46.73
81stof 170Grok 4.5Accuracy46.43
82ndof 170Claude Sonnet 4Accuracy46.11
84thof 170Claude Opus 4.1Accuracy45.86
86thof 170DeepSeek V4 FlashAccuracy (max)45.63
87thof 170Kimi K2.5Accuracy45.6
88thof 170Claude Haiku 4.5Accuracy45.4
92ndof 170Claude Opus 4Accuracy43.72
94thof 170o3-miniAccuracy (high)43.7
95thof 170Nemotron 3 UltraAccuracy43.45
96thof 170Mercury 2Accuracy43.2
97thof 170Grok 4 FastAccuracy42.86
98thof 170Kimi K2 ThinkingAccuracy42.79
99thof 170Grok 3 MiniAccuracy (high)42.58
103rdof 170DeepSeek-R1 (0528)Accuracy41.63
104thof 170Qwen3-Coder-480B-A35B-InstructAccuracy41.17
105thof 170Qwen3 235B A22B Thinking 2507Accuracy41.04
106thof 170Gemini 2.5 FlashAccuracy40.95
108thof 170GPT OSS 20BAccuracy (high)40.93
110thof 170Claude 3.5 SonnetAccuracy39.97
111thof 170GPT-5 ChatAccuracy39.77
112thof 170Qwen3.5-27BAccuracy39.53
115thof 170Kimi K2 InstructAccuracy39.36
116thof 170Kimi K2 0711Accuracy39.36
117thof 170GPT-4.1 fine-tunedAccuracy39.04
118thof 170Gemini 3.5 Flash-LiteAccuracy (high)39
119thof 170Qwen3-235B-A22B-Instruct-2507Accuracy38.7
122ndof 170GPT-5 nanoAccuracy (high)38.06
123rdof 170Nemotron 3 SuperAccuracy38.01
126thof 170GPT-4.1 miniAccuracy37.61
127thof 170DeepSeek-V3.1Accuracy37.5
129thof 170Qwen3 235B A22BAccuracy37.28
130thof 170Grok 3Accuracy37.24
131stof 170MiniMax M2.7Accuracy36.95
134thof 170DeepSeek-R1Accuracy36.49
135thof 170o1-miniAccuracy (medium)36.32
136thof 170DeepSeek-V3 0324Accuracy36.08
137thof 170Gemini 2.5 Flash-LiteAccuracy35.22
139thof 170Gemma 4 26B A4BAccuracy35.17
140thof 170grok-code-fast-1Accuracy35.06
141stof 170Qwen3.6 35B A3BAccuracy34.49
142ndof 170Qwen3 Coder NextAccuracy34.4
144thof 170InklingAccuracy (high)32.29
146thof 170Claude Haiku 3.5Accuracy30.73
147thof 170Qwen3 30B A3BAccuracy29.75
149thof 170gemini-2.0-flash-001Accuracy25.77
150thof 170GPT-4o (2024-11-20)Accuracy25.12
152ndof 170Llama-4-Maverick-17B-128E-InstructAccuracy24.47
155thof 170Meta-Llama-3.1-405B-InstructAccuracy21.38
156thof 170Claude 3 OpusAccuracy19.22
157thof 170GPT-4.1 nanoAccuracy18.98
158thof 170GPT-4 TurboAccuracy18.01
159thof 170Qwen2.5 72B InstructAccuracy15.97
160thof 170Llama 3.3 70B InstructAccuracy14.44
161stof 170GPT-4 (0613)Accuracy12.36
162ndof 170GPT-4o-mini (2024-07-18)Accuracy11.76
163rdof 170Qwen2-72B-InstructAccuracy11.3
164thof 170Claude 3 SonnetAccuracy10.16
165thof 170Claude 3 HaikuAccuracy9.84
166thof 170Meta-Llama-3.1-70B-InstructAccuracy8.97
167thof 170Claude 2.1Accuracy7.06
168thof 170GPT-3.5 TurboAccuracy3.48
169thof 170Mixtral-8x22B-Instruct-v0.1Accuracy3.17
170thof 170Llama 3.1 8B InstructAccuracy1.73

Read from the board on 2026-09-16