Knowledge & reasoning

SimpleQA leaderboard

Name: benchgap: LLM benchmark scores, measured and estimated
Creator: benchgap

As of 2026-10-10, the highest measured score on SimpleQA is 54.3% by Qwen3 235B 2507. 101 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	GPT-6 Astra	76.9%	estimated ± 10.9 pp, low confidence
2	Sakana Fugu	76.1%	estimated ± 10.9 pp, low confidence
3	Sakana Fugu-Ultra	76.1%	estimated ± 10.9 pp, low confidence
4	GPT-5.6 Sol	74.7%	estimated ± 10.9 pp, low confidence
5	Claude Opus 4.7 (Adaptive)	74.0%	estimated ± 10.9 pp, low confidence
6	Claude Mythos 5	73.8%	estimated ± 10.9 pp, low confidence
7	Claude Opus 4.8	73.0%	estimated ± 10.9 pp, low confidence
8	GPT-5.5	73.0%	estimated ± 10.9 pp, low confidence
9	Kimi K3	72.8%	estimated ± 10.9 pp, low confidence
10	GPT-5.6 Terra	71.8%	estimated ± 10.9 pp, low confidence
11	GPT-5.4	71.7%	estimated ± 10.9 pp, low confidence
12	Ornith-1.5-397B	71.7%	estimated ± 10.9 pp, low confidence
13	Qwen3.8 Max	71.3%	estimated ± 10.9 pp, low confidence
14	GPT-5.2	71.0%	estimated ± 10.9 pp, low confidence
15	Qwen3.7 Max	71.0%	estimated ± 10.9 pp, low confidence
16	GPT-5.6 Luna	70.8%	estimated ± 10.9 pp, low confidence
17	Hy4 preview	70.8%	estimated ± 10.9 pp, low confidence
18	Gemini 3.5 Flash	70.6%	estimated ± 10.9 pp, low confidence
19	Qwen3.8-Flash-Next	69.8%	estimated ± 10.9 pp, low confidence
20	Claude Opus 4.6	69.1%	estimated ± 10.9 pp, low confidence
21	GLM-5.2	68.9%	estimated ± 10.9 pp, low confidence
22	Qwen3.8-Omni-Flash	68.6%	estimated ± 10.9 pp, low confidence
23	DeepSeek V4.1 Flash	68.4%	estimated ± 10.9 pp, low confidence
24	Kimi K2.6	67.7%	estimated ± 10.9 pp, low confidence
25	Qwen3.6 Plus	67.5%	estimated ± 10.9 pp, low confidence
26	Qwen3.7 Plus	67.3%	estimated ± 10.9 pp, low confidence
27	DeepSeek V4 Pro 0813	67.0%	estimated ± 10.9 pp, low confidence
28	Grok 4.3	67.0%	estimated ± 10.9 pp, low confidence
29	Claude Sonnet 4.6	66.6%	estimated ± 10.9 pp, low confidence
30	Interfaze Beta	66.6%	estimated ± 10.9 pp, low confidence
31	Inkling-Small	65.9%	estimated ± 10.9 pp, low confidence
32	Ornith-1.5-35B-A3B	65.4%	estimated ± 10.9 pp, low confidence
33	Qwen3.8-27B	65.4%	estimated ± 10.9 pp, low confidence
34	Qwen3.5 397B	64.0%	estimated ± 10.9 pp, low confidence
35	DeepSeek V4 Flash 0731	63.4%	estimated ± 10.9 pp, low confidence
36	GPT-5.4 mini	63.2%	estimated ± 10.9 pp, low confidence
37	Inkling	63.1%	estimated ± 10.9 pp, low confidence
38	Qwen3.6-27B	62.9%	estimated ± 10.9 pp, low confidence
39	Kimi K2.5	62.5%	estimated ± 10.9 pp, low confidence
40	Kimi K2.5 (Reasoning)	62.5%	estimated ± 10.9 pp, low confidence
41	Hy3 Preview	61.8%	estimated ± 10.9 pp, low confidence
42	Claude Opus 4.5	61.4%	estimated ± 10.9 pp, low confidence
43	Nemotron 3 Ultra	61.4%	estimated ± 10.9 pp, low confidence
44	Qwen3.5-122B-A10B	60.7%	estimated ± 10.9 pp, low confidence
45	Ornith-1.5-9B	60.4%	estimated ± 10.9 pp, low confidence
46	GLM-5	59.6%	estimated ± 10.9 pp, low confidence
47	Qwen3.6-35B-A3B	59.6%	estimated ± 10.9 pp, low confidence
48	Ternary Bonsai 2 27B	59.2%	estimated ± 10.9 pp, low confidence
49	GLM-4.7	59.1%	estimated ± 10.9 pp, low confidence
50	Qwen3.5-27B	58.7%	estimated ± 10.9 pp, low confidence
51	Ling 3.0 Flash	57.8%	estimated ± 10.9 pp, low confidence
52	Gemma 4 31B	56.6%	estimated ± 10.9 pp, low confidence
53	MAI-Thinking-1	56.4%	estimated ± 10.9 pp, low confidence
54	Qwen3.5-35B-A3B	56.4%	estimated ± 10.9 pp, low confidence
55	Ling 3.0 Flash FP8	56.0%	estimated ± 10.9 pp, low confidence
56	MiMo-V2-Flash	55.5%	estimated ± 10.9 pp, low confidence
57	Claude Sonnet 4.5	54.9%	estimated ± 10.9 pp, low confidence
58	Qwen3 235B 2507	54.3%	measured
59	Gemini 2.5 Pro	54.2%	estimated ± 10.9 pp, low confidence
60	GPT-5.4 nano	53.9%	estimated ± 10.9 pp, low confidence
61	Gemma 4 26B A4B	53.0%	estimated ± 10.9 pp, low confidence
62	Qwen3 235B 2507 (Reasoning)	50.8%	estimated ± 10.9 pp, low confidence
63	Grok 3 Mini	49.4%	estimated ± 10.9 pp, low confidence
64	DeepSeek V3.1 (Reasoning)	49.1%	estimated ± 10.9 pp, low confidence
65	GPT-OSS 120B	49.1%	estimated ± 10.9 pp, low confidence
66	Nemotron 3 Super 100B	47.6%	estimated ± 10.9 pp, low confidence
67	o1-pro	47.2%	estimated ± 10.9 pp, low confidence
68	Gemma 4 12B	46.8%	estimated ± 10.9 pp, low confidence
69	o3-mini	44.1%	estimated ± 10.9 pp, low confidence
70	Grok 3 [Beta]	43.6%	measured
71	Nemotron Ultra 253B	42.2%	estimated ± 10.9 pp, low confidence
72	o1	41.6%	estimated ± 10.9 pp, low confidence
73	Nemotron 3.5 Lightning 30B A3B NVFP4	41.4%	estimated ± 10.9 pp, low confidence
74	GLM-4.7-Flash	40.8%	estimated ± 10.9 pp, low confidence
75	DeepSeek V3.1	40.4%	estimated ± 10.9 pp, low confidence
76	Nemotron 3 Nano 30B	37.5%	estimated ± 10.9 pp, low confidence
77	Nemotron 3 Nano Omni 30B A3B	36.3%	estimated ± 10.9 pp, low confidence
78	GPT-OSS 20B	35.3%	estimated ± 10.9 pp, low confidence
79	Ministral 3 14B (Reasoning)	34.9%	estimated ± 10.9 pp, low confidence
80	ZAYA1-8B	34.6%	estimated ± 10.9 pp, low confidence
81	Kimi K2	31.0%	measured
82	DeepSeek-R1	30.1%	measured
83	Granite 4.2 30B	28.8%	estimated ± 10.9 pp, low confidence
84	GPT-4.1	28.7%	estimated ± 10.9 pp, low confidence
85	GPT-4.1 mini	26.4%	estimated ± 10.9 pp, low confidence
86	Granite 4.2 8B	26.4%	estimated ± 10.9 pp, low confidence
87	DeepSeek V3	24.9%	measured
88	Claude 3.5 Sonnet	22.2%	estimated ± 10.9 pp, low confidence
89	Ling 2.6 Flash	21.9%	estimated ± 10.9 pp, low confidence
90	Gemma 4 E4B	21.6%	estimated ± 10.9 pp, low confidence
91	Mellum2-12B-A2.5B-Thinking	21.0%	estimated ± 10.9 pp, low confidence
92	ZAYA1-74B-Preview	20.8%	estimated ± 10.9 pp, low confidence
93	Phi-4	20.0%	estimated ± 10.9 pp, low confidence
94	Granite 4.2 3B	19.3%	estimated ± 10.9 pp, low confidence
95	GPT-4.1 nano	17.3%	estimated ± 10.9 pp, low confidence
96	Qwen2.5-VL-32B	16.1%	estimated ± 10.9 pp, low confidence
97	Gemma 4 E2B	15.5%	estimated ± 10.9 pp, low confidence
98	Soofi S 30B-A3B	15.5%	estimated ± 10.9 pp, low confidence
99	Mellum2-12B-A2.5B-Instruct	15.1%	estimated ± 10.9 pp, low confidence
100	LFM2.5-350M	14.4%	estimated ± 10.9 pp, low confidence
101	Granite-4.0-1B	14.4%	estimated ± 10.9 pp, low confidence
102	Granite-4.0-H-1B	14.4%	estimated ± 10.9 pp, low confidence
103	Granite-4.0-H-350M	14.3%	estimated ± 10.9 pp, low confidence
104	LFM2.5-VL-450M	14.3%	estimated ± 10.9 pp, low confidence
105	LFM2.5-230M	14.3%	estimated ± 10.9 pp, low confidence
106	Granite-4.0-350M	14.3%	estimated ± 10.9 pp, low confidence

SimpleQA leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General