Inference
Epoch AI

SimpleQA Verified

Evaluated by Epoch AI on their own harness.

ModelsModels with a result.67
Top resultBest result on this test.75.6%GPT-6 Astra
Top-3 spreadPoints from first to third.4.8 pts
YearYear of release.N/A

Result and price

020406080100$0.3$1$3$10$30$100$300RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

67 results
#ModelResultThe score from the source.
01GPT-6 Astramax effort75.6% ±1.4
02Gemini 3.1 Prohigh effort73.5% ±1.4
03Claude Fable 5.1max effort70.8% ±1.4
04Claude Fable 5xhigh effort70.7% ±1.4
05GPT-5.6 Solmax effort69.7% ±1.5
05Gemini 3.8 Flashhigh effort69.7% ±1.5
07Gemini 3.7 Flashhigh effort69.2% ±1.5
08Gemini 3 Flashhigh effort66.8% ±1.5
09Gemini 3.5 Flashhigh effort66.2% ±1.5
09Gemini 3.6 Flashhigh effort66.2% ±1.5
11GPT-5.5xhigh effort63.0% ±1.5
12Muse Spark 1.2xhigh effort60.3% ±1.6
13Claude Opus 5max effort59.9% ±1.6
14Muse Spark 1.157.8% ±1.6
15Qwen3.7 Max55.8% ±1.6
16Claude Opus 4.8max effort53.0% ±1.6
17Qwen 3.6 Max (preview)52.0% ±1.6
18Claude Opus 4.7xhigh effort51.7% ±1.6
19GPT-5high effort50.1% ±1.6
20o3high effort49.4% ±1.6
21Grok 4.6xhigh effort48.9% ±1.6
22Qwen3 Max48.7% ±1.6
23Grok 4.5high effort48.3% ±1.6
24GPT-5.1high effort48.0% ±1.6
25Qwen3.8 Max (0902)xhigh effort47.3% ±1.6
26Claude Opus 4.6max effort47.0% ±1.6
27DeepSeek V4 Pro 0423max effort47.0% ±1.6
28GPT-5.4 Proxhigh effort46.3% ±1.6
29Qwen3.8 Maxxhigh effort45.8% ±1.6
30Claude Opus 4.545.7% ±1.6
31GPT-5.4xhigh effort45.1% ±1.6
32Qwen3.6 Plus44.1% ±1.6
33GPT-5.6 Terramax effort43.2% ±1.6
34o1high effort41.1% ±1.6
35GPT-5.6 Lunamax effort41.0% ±1.6
35GLM-5.3max effort41.0% ±1.6
37Qwen3-235B-A22B (Jul 2025)40.4% ±1.6
38Inklingxhigh effort40.3% ±1.6
39GPT-5.2xhigh effort37.1% ±1.5
40Kimi K2.7 Code36.5% ±1.5
41Kimi K2.534.3% ±1.5
42GLM-5.2max effort34.2% ±1.5
43GLM-5.134.0% ±1.5
44Claude Sonnet 5max effort33.7% ±1.5
45Grok 4.3 Betahigh effort33.2% ±1.5
46Claude Sonnet 4.6max effort32.8% ±1.5
47GLM-4.732.2% ±1.5
48GPT-4.131.1% ±1.5
49Grok 4.2030.2% ±1.5
50GPT-5.4 minihigh effort29.4% ±1.4
51Claude Sonnet 4.527.2% ±1.5
52GPT-4o26.0% ±1.4
53Qwen 3.5 Plus (hosted 397B-A17B)25.4% ±1.4
54GPT-5 minihigh effort21.6% ±1.3
55Qwen 3.5 Flash (hosted 35B-A3B)20.3% ±1.3
56Inkling-Smallxhigh effort19.1% ±1.2
57o4-mini (high)high effort18.8% ±1.2
58Qwen3.6 Flash15.9% ±1.2
59o3-minihigh effort15.3% ±1.1
60Claude Haiku 4.512.9% ±1.0
61GPT-4.1 mini12.7% ±1.1
62Claude 3 Opus12.6% ±1.0
63GPT-5.4 nanohigh effort11.7% ±1.0
63GPT-5 nanohigh effort11.7% ±1.0
65Gemma 4 31B10.4% ±1.0
66GPT-4o mini8.3% ±0.9
67GPT-4.1 nano6.0% ±0.8