Agentic · tools

WideResearch leaderboard

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

As of 2026-10-07, the highest measured score on WideResearch is 83.9% by Hy4 preview. 155 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#	Model	Score	Source
1	Claude Fable 5	100.0%	estimated ± 9.1 pp, low confidence
2	Gemini 4 Argon	100.0%	estimated ± 9.1 pp, low confidence
3	GPT-6.1 Sol	100.0%	estimated ± 9.1 pp, low confidence
4	GPT-6 Sol	100.0%	estimated ± 9.1 pp, low confidence
5	Grok 4.6	100.0%	estimated ± 9.1 pp, low confidence
6	Grok 4.7	100.0%	estimated ± 9.1 pp, low confidence
7	Qwen3.8 Max Preview	100.0%	estimated ± 9.1 pp, low confidence
8	Muse Spark 1.2	98.8%	estimated ± 9.1 pp, low confidence
9	GPT-6 Luna	97.1%	estimated ± 9.1 pp, low confidence
10	Gemini 3.8 Flash	97.0%	estimated ± 9.1 pp, low confidence
11	Mistral Large 4	96.5%	estimated ± 9.1 pp, low confidence
12	Gemini 3.7 Flash	95.1%	estimated ± 9.1 pp, low confidence
13	Grok 4.5	95.0%	estimated ± 9.1 pp, low confidence
14	Gemini 3.6 Flash	90.3%	estimated ± 9.1 pp, low confidence
15	Claude Opus 5.5	87.9%	estimated ± 3.4 pp, low confidence
16	Claude Opus 5	86.5%	estimated ± 3.4 pp, low confidence
17	Claude Sonnet 5.5	86.4%	estimated ± 3.4 pp, low confidence
18	DeepSeek V4.1 Flash	86.2%	estimated ± 3.4 pp, low confidence
19	Apodex 1.1 Mini	86.0%	estimated ± 9.1 pp, low confidence
20	Claude Fable 5.1	85.6%	estimated ± 3.9 pp, low confidence
21	GLM-5.3	85.5%	estimated ± 3.4 pp, low confidence
22	Quasar 438B	85.4%	estimated ± 9.1 pp, low confidence
23	Gemini 3.5 Flash	84.7%	estimated ± 5.6 pp, low confidence
24	Muse Spark 1.3	84.5%	estimated ± 1.2 pp, low confidence
25	Ling 3.0 Flash VL	84.4%	estimated ± 9.1 pp, low confidence
26	DeepSeek V4 Pro 0813	84.2%	estimated ± 3.4 pp, low confidence
27	Hy4 preview	83.9%	measured
28	MiMo-V2.6-Pro	83.9%	estimated ± 1.2 pp, low confidence
29	MiMo-V2.6-Flash	83.7%	estimated ± 1.2 pp, medium confidence
30	Step 5 Preview	83.2%	estimated ± 1.2 pp, medium confidence
31	Claude Haiku 5.5	82.8%	estimated ± 3.4 pp, low confidence
32	Claude Sonnet 5	82.8%	estimated ± 3.4 pp, low confidence
33	GPT-5.6 Sol	82.8%	estimated ± 4.6 pp, high confidence
34	GPT-6 Astra	82.7%	estimated ± 3.4 pp, low confidence
35	Ling 3.1 Flash	82.7%	estimated ± 1.5 pp, medium confidence
36	Qwen3.8-Flash-Next	82.5%	estimated ± 1.2 pp, medium confidence
37	Muse Spark 1.1	82.3%	estimated ± 1.2 pp, medium confidence
38	Apodex 1.1	82.1%	estimated ± 3.4 pp, medium confidence
39	GPT-5.5 Pro	82.0%	estimated ± 4.6 pp, high confidence
40	Atria Dawn Preview	81.9%	measured
41	Qwen3.8 Max	81.9%	measured
42	Kimi K3	81.9%	estimated ± 1.2 pp, medium confidence
43	GPT-5.4 Pro	81.7%	estimated ± 4.6 pp, high confidence
44	GLM-5.3-Flash	81.6%	estimated ± 3.4 pp, medium confidence
45	Claude Mythos 5	81.2%	estimated ± 4.6 pp, high confidence
46	GPT-5.6 Terra	81.0%	estimated ± 4.6 pp, high confidence
47	Kimi K2.6	80.8%	measured
48	Ornith-1.5-397B	80.8%	measured
49	Qwen3.7 Max	80.6%	estimated ± 3.4 pp, medium confidence
50	Claude Opus 4.7 (Adaptive)	80.4%	estimated ± 1.2 pp, medium confidence
51	GPT-5.5	79.7%	estimated ± 1.2 pp, medium confidence
52	Claude Opus 4.8	79.7%	estimated ± 4.6 pp, high confidence
53	GLM-5.2	79.4%	estimated ± 5.6 pp, low confidence
54	Claude Opus 4.7	79.4%	estimated ± 5.9 pp, low confidence
55	MiniMax M3	79.3%	estimated ± 4.6 pp, high confidence
56	Hy3	79.2%	estimated ± 9.1 pp, low confidence
57	GPT-5.6 Luna	79.2%	estimated ± 4.6 pp, high confidence
58	MiMo-V2.5-Pro	78.9%	estimated ± 5.9 pp, low confidence
59	dots3-note Preview	78.9%	measured
60	GPT-5.4	78.8%	estimated ± 1.2 pp, medium confidence
61	Claude Sonnet 4.6	78.4%	estimated ± 1.2 pp, medium confidence
62	Claude Opus 4.6	78.4%	estimated ± 1.2 pp, medium confidence
63	MiniMax M2.7	78.2%	estimated ± 5.6 pp, low confidence
64	GPT-5.2	77.8%	estimated ± 1.2 pp, medium confidence
65	Kimi K2.7 Code	77.8%	estimated ± 9.1 pp, low confidence
66	GPT-5.3 Codex	77.7%	estimated ± 1.2 pp, medium confidence
67	Qwen3.8-27B	77.6%	estimated ± 1.2 pp, medium confidence
68	Gemini 3.1 Pro	77.3%	estimated ± 5.9 pp, low confidence
69	Agents-A1	76.8%	estimated ± 3.4 pp, medium confidence
70	Beam	76.6%	estimated ± 4.6 pp, high confidence
71	Inkling	76.5%	estimated ± 4.6 pp, high confidence
72	Qwen3.6-27B	76.5%	estimated ± 5.9 pp, low confidence
73	Step 3.7 Flash	76.5%	estimated ± 3.4 pp, medium confidence
74	Claude Opus 4.5	76.4%	measured
75	Claude Sonnet 4.5	76.4%	estimated ± 1.2 pp, medium confidence
76	GPT-5.1-Codex	76.1%	estimated ± 1.2 pp, medium confidence
77	GPT-5.4 mini	76.0%	estimated ± 5.6 pp, low confidence
78	GPT-5.2-Codex	76.0%	estimated ± 1.2 pp, medium confidence
79	Muse Spark	75.7%	estimated ± 9.1 pp, low confidence
80	Claude 4.1 Opus	75.1%	estimated ± 1.2 pp, medium confidence
81	DeepSeek V4 Flash 0731	75.0%	estimated ± 3.4 pp, medium confidence
82	Gemini 3.5 Flash-Lite	74.8%	estimated ± 9.1 pp, low confidence
83	Qwen3.5 Plus	74.4%	estimated ± 1.2 pp, medium confidence
84	Claude 4 Sonnet	74.4%	estimated ± 1.2 pp, medium confidence
85	Qwen3.6 Plus	74.3%	measured
86	Qwen3.5 397B	74.0%	measured
87	Claude Haiku 4.5	73.8%	estimated ± 1.2 pp, medium confidence
88	Ling 3.0 Flash	73.6%	measured
89	A.X K2	73.5%	estimated ± 9.1 pp, low confidence
90	Muse Glimmer 30B	73.0%	estimated ± 1.5 pp, low confidence
91	Gemini 3 Flash	72.8%	estimated ± 1.2 pp, medium confidence
92	Gemini 3 Pro	72.8%	estimated ± 1.2 pp, medium confidence
93	Grok Build 0.1	72.8%	estimated ± 5.9 pp, low confidence
94	Kimi K2.5	72.7%	measured
95	Ling 3.0 Flash FP8	72.6%	estimated ± 9.1 pp, low confidence
96	Inkling-Small	72.4%	estimated ± 3.9 pp, medium confidence
97	GPT-5 (high)	72.2%	estimated ± 1.2 pp, low confidence
98	GLM-5.1	72.0%	estimated ± 4.6 pp, high confidence
99	Agents-A1-4B	71.4%	estimated ± 4.6 pp, high confidence
100	Pokee-Isaac 28B	71.4%	estimated ± 12.4 pp, low confidence
101	GPT-5.4 nano	71.3%	estimated ± 5.6 pp, low confidence
102	Grok 4.1 Fast	70.9%	estimated ± 5.9 pp, low confidence
103	MiMo-V2.5	70.4%	estimated ± 5.9 pp, low confidence
104	GLM-5	69.8%	measured
105	Qwen3.5-122B-A10B	69.8%	estimated ± 4.6 pp, high confidence
106	Laguna S 2.1	69.1%	estimated ± 3.9 pp, medium confidence
107	Nemotron 3 Ultra	68.8%	estimated ± 3.4 pp, medium confidence
108	Qwen3.5-27B	68.2%	estimated ± 4.6 pp, high confidence
109	Qwen3.5-35B-A3B	68.2%	estimated ± 4.6 pp, high confidence
110	Kimi K2.5 (Reasoning)	67.9%	estimated ± 4.6 pp, high confidence
111	Ornith-1.5-35B-A3B	67.8%	measured
112	Grok 4.3	66.2%	estimated ± 5.9 pp, low confidence
113	Qwen3 Max	66.0%	estimated ± 5.9 pp, low confidence
114	Grok 4	63.4%	estimated ± 5.9 pp, low confidence
115	Gemini 2.5 Pro	62.8%	estimated ± 5.9 pp, low confidence
116	GLM-4.7	62.5%	estimated ± 4.6 pp, medium confidence
117	Qwen3.7 Plus	61.5%	estimated ± 9.1 pp, low confidence
118	GPT-5.1	61.2%	estimated ± 5.9 pp, low confidence
119	Solar Pro 4	60.6%	estimated ± 4.6 pp, medium confidence
120	LongCat-Flash-Lite-Sparse	60.2%	estimated ± 4.6 pp, medium confidence
121	Qwen3.6-35B-A3B	60.1%	measured
122	Ornith-1.5-9B	59.5%	measured
123	MiniCPM5-2B	57.4%	estimated ± 9.1 pp, low confidence
124	Mistral Medium 3.5 128B	56.0%	estimated ± 5.9 pp, low confidence
125	Gemini 3.1 Flash-Lite	54.2%	estimated ± 5.9 pp, low confidence
126	Grok 4.20	54.0%	estimated ± 5.9 pp, low confidence
127	MiMo-V2-Flash	51.9%	estimated ± 9.1 pp, low confidence
128	Nemotron 3.5 Lightning 30B A3B NVFP4	50.8%	estimated ± 4.6 pp, medium confidence
129	Hy3 Preview	49.7%	estimated ± 5.9 pp, low confidence
130	MiMo-V2-Pro	48.9%	estimated ± 5.9 pp, low confidence
131	Granite 4.2 30B	47.8%	estimated ± 9.1 pp, low confidence
132	Gemma 4 26B A4B	47.0%	estimated ± 9.1 pp, low confidence
133	Ling 3.0 Tiny	46.7%	estimated ± 9.1 pp, low confidence
134	Gemma 4 31B	44.2%	estimated ± 5.9 pp, low confidence
135	Command A+	43.1%	estimated ± 9.1 pp, low confidence
136	Celeris-1	41.6%	estimated ± 9.1 pp, low confidence
137	DeepSeek V3	41.6%	estimated ± 9.1 pp, low confidence
138	DeepSeek V3 0324	41.6%	estimated ± 9.1 pp, low confidence
139	Gemma 3 27B	41.6%	estimated ± 9.1 pp, low confidence
140	Gemma 4 12B	41.6%	estimated ± 9.1 pp, low confidence
141	Gemma 4 E2B	41.6%	estimated ± 9.1 pp, low confidence
142	Gemma 4 E4B	41.6%	estimated ± 9.1 pp, low confidence
143	GPT-4.1 mini	41.6%	estimated ± 9.1 pp, low confidence
144	GPT-4.1 nano	41.6%	estimated ± 9.1 pp, low confidence
145	GPT-4o	41.6%	estimated ± 9.1 pp, low confidence
146	GPT-4o mini	41.6%	estimated ± 9.1 pp, low confidence
147	GPT-OSS 20B	41.6%	estimated ± 9.1 pp, low confidence
148	Granite 4.2 3B	41.6%	estimated ± 9.1 pp, low confidence
149	Granite 4.2 8B	41.6%	estimated ± 9.1 pp, low confidence
150	K-Exaone	41.6%	estimated ± 9.1 pp, low confidence
151	LFM2.5-2.6B	41.6%	estimated ± 9.1 pp, low confidence
152	Ling 2.6 Flash	41.6%	estimated ± 9.1 pp, low confidence
153	Llama 4 Maverick	41.6%	estimated ± 9.1 pp, low confidence
154	Llama 4 Scout	41.6%	estimated ± 9.1 pp, low confidence
155	Mercury 2.5	41.6%	estimated ± 9.1 pp, low confidence
156	Mistral Large 3	41.6%	estimated ± 9.1 pp, low confidence
157	Mistral Small 4	41.6%	estimated ± 9.1 pp, low confidence
158	Mistral Small 4 (Reasoning)	41.6%	estimated ± 9.1 pp, low confidence
159	Nemotron 3 Nano 30B	41.6%	estimated ± 9.1 pp, low confidence
160	Nemotron 3 Nano Omni 30B A3B	41.6%	estimated ± 9.1 pp, low confidence
161	Nemotron 3 Super 100B	41.6%	estimated ± 9.1 pp, low confidence
162	North Mini Code	41.6%	estimated ± 9.1 pp, low confidence
163	Solar Pro 3	41.6%	estimated ± 9.1 pp, low confidence
164	Trinity-Large-Preview	41.6%	estimated ± 9.1 pp, low confidence
165	Ultravox v0.6 Llama 3.3 70B	41.6%	estimated ± 9.1 pp, low confidence
166	Trinity-Large-Thinking	34.6%	estimated ± 5.9 pp, low confidence
167	GLM-5V-Turbo	28.1%	estimated ± 5.9 pp, low confidence
168	GPT-OSS 120B	24.0%	estimated ± 5.9 pp, low confidence
169	DeepSeek V3.2	23.9%	estimated ± 5.9 pp, low confidence
170	GPT-4.1	12.3%	estimated ± 5.9 pp, low confidence

WideResearch leaderboard

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General