Inference
EN
Epoch AI

SimpleQA Verified

Evaluated by Epoch AI on their own harness.

ModelsModels with a result.91
Top resultBest result on this benchmark.94.0%Claude Fable 5.2
Top-3 spreadPoints from first to third.18.0 pts
YearYear of release.—

Result and price

79 of 91 results

Results

91 results
#ModelResultThe score from the source.
01Claude Fable 5.2max effort94.0% ±0.8
02Gemini 4 Carbonhigh effort77.7% ±1.3
03Gemini 4 Argonhigh effort76.0% ±1.4
04GPT-6 Astramax effort75.6% ±1.4
05GPT-6.1 Solmax effort73.9% ±1.4
06Gemini 3.1 Prohigh effortAbove fitFar above this model's index.73.5% ±1.4
07Claude Opus 5.5max effort72.2% ±1.4
08Claude Fable 5.1max effort70.8% ±1.4
09Claude Fable 5xhigh effort70.7% ±1.4
10GPT-5.6 Solmax effort69.7% ±1.5
10Gemini 3.8 Flashhigh effort69.7% ±1.5
12Gemini 3.7 Flashhigh effort69.2% ±1.5
13Gemini 3 Flashhigh effortAbove fitFar above this model's index.66.8% ±1.5
14Gemini 3.5 Flashhigh effort66.2% ±1.5
14Gemini 3.6 Flashhigh effort66.2% ±1.5
16GPT-5.5xhigh effort63.0% ±1.5
17GPT-6 Solmax effort60.7% ±1.5
18Muse Spark 1.2xhigh effort60.3% ±1.6
19Claude Opus 5max effort59.9% ±1.6
20Muse Spark 1.157.8% ±1.6
21GPT-6.1 Terramax effort57.3% ±1.6
22Muse Spark Openxhigh effort56.6% ±1.6
23Grok 4.7xhigh effort56.0% ±1.6
24Qwen3.7 Max55.8% ±1.6
25Claude Opus 4.8max effort53.0% ±1.6
26DeepSeek V4 Pro 0813max effort52.9% ±1.6
27Qwen 3.6 Max (preview)52.0% ±1.6
28Claude Opus 4.7xhigh effort51.7% ±1.6
29Kimi K3max effort50.6% ±1.6
30GPT-5high effort50.1% ±1.6
31o3high effort49.4% ±1.6
32Grok 4.6xhigh effort48.9% ±1.6
33Qwen3 MaxAbove fitFar above this model's index.48.7% ±1.6
34Grok 4.5high effort48.3% ±1.6
35GPT-5.1high effort48.0% ±1.6
36Qwen3.8 Max (0902)xhigh effort47.3% ±1.6
37Claude Opus 4.6max effort47.0% ±1.6
37DeepSeek V4 Pro 0423max effort47.0% ±1.6
39Claude Sonnet 5.5max effort46.5% ±1.6
40GPT-5.4 Proxhigh effort46.3% ±1.6
41Qwen3.8 Maxxhigh effort45.8% ±1.6
42Claude Opus 4.545.7% ±1.6
43GPT-5.4xhigh effort45.1% ±1.6
44Doubao Seed 2.1 Prohigh effort44.7% ±1.6
45Qwen3.6 Plus44.1% ±1.6
46GPT-5.6 Terramax effort43.2% ±1.6
47DeepSeek V4.1 Flash 1017max effort42.9% ±1.6
48Doubao Seed Evolving (Nov 2026)high effort41.9% ±1.6
49GPT-6 Lunamax effort41.4% ±1.6
50o1high effort41.1% ±1.6
51GLM-5.3max effort41.0% ±1.6
51GPT-5.6 Lunamax effort41.0% ±1.6
53Qwen3-235B-A22B (Jul 2025)40.4% ±1.6
54Inklingxhigh effort40.3% ±1.6
55Doubao-Seed-2.0 Prohigh effort39.4% ±1.5
56GPT-5.2xhigh effort37.1% ±1.5
57Kimi K2.7 Code36.5% ±1.5
58Kimi K2.634.9% ±1.5
59Kimi K2.534.3% ±1.5
60GLM-5.2max effort34.2% ±1.5
61GLM-5.134.0% ±1.5
62Claude Sonnet 5max effort33.7% ±1.5
63DeepSeek V4 Flash 0731max effort33.6% ±1.5
64Grok 4.3 Betahigh effort33.2% ±1.5
65Claude Sonnet 4.6max effort32.8% ±1.5
66GLM-4.732.2% ±1.5
67Doubao Seed 2.0 Lite (Apr 2026)high effort31.5% ±1.5
68GPT-4.131.1% ±1.5
69Grok 4.2030.2% ±1.5
70Doubao Seed 2.1 Turbohigh effort29.7% ±1.4
71GPT-5.4 minihigh effort29.4% ±1.4
72Claude Sonnet 4.527.2% ±1.5
73GPT-4o26.0% ±1.4
74Qwen 3.5 Plus (hosted 397B-A17B)25.4% ±1.4
75Claude Haiku 5.5max effort23.8% ±1.3
76GPT-5 minihigh effort21.6% ±1.3
77Mistral Large 4high effort21.2% ±1.3
78Qwen 3.5 Flash (hosted 35B-A3B)20.3% ±1.3
79Inkling-Smallxhigh effort19.1% ±1.2
80o4-minihigh effort18.8% ±1.2
81Doubao Seed 1.617.4% ±1.2
82Qwen3.6 Flash15.9% ±1.2
83o3-minihigh effort15.3% ±1.1
84Claude Haiku 4.512.9% ±1.0
85GPT-4.1 mini12.7% ±1.1
86Claude 3 Opus12.6% ±1.0
87GPT-5.4 nanohigh effort11.7% ±1.0
87GPT-5 nanohigh effort11.7% ±1.0
89Gemma 4 31B10.4% ±1.0
90GPT-4o mini8.3% ±0.9
91GPT-4.1 nano6.0% ±0.8