benchgap
General

FACTS Multimodal leaderboard

As of 2026-11-18, the highest measured score on FACTS Multimodal is 60.2% by Gemini 4 Carbon. 102 more models have estimated scores, calibrated from the benchmarks they were measured on.

Measured scores: benchlm.ai.

#ModelScoreSource
1Gemini 4 Carbon60.2%measured
2Gemini 3.8 Flash57.6%measured
3Claude Opus 5.556.4%measured
4GPT-6 Astra55.2%measured
5Claude Fable 5.254.2%estimated ± 4.0 pp, high confidence
6Gemini 4 Argon54.2%measured
7Gemini 3.6 Flash53.7%measured
8Gemini 3 Pro52.7%estimated ± 4.0 pp, high confidence
9Claude Opus 552.6%estimated ± 4.0 pp, high confidence
10GPT-6.1 Sol52.6%measured
11Gemini 3.7 Flash52.5%estimated ± 4.0 pp, high confidence
12GPT-6.2 Astra52.4%estimated ± 2.7 pp, medium confidence
13GPT-6.2 Sol52.4%estimated ± 2.7 pp, medium confidence
14GPT-6 Cyber52.4%estimated ± 2.7 pp, medium confidence
15Muse Wonder52.2%estimated ± 4.0 pp, high confidence
16Claude Fable 551.8%estimated ± 4.0 pp, high confidence
17Gemini 3.5 Flash51.7%estimated ± 4.0 pp, high confidence
18GPT-5.5 Pro51.6%estimated ± 5.6 pp, medium confidence
19Qwen3.8 Max51.5%measured
20Claude Opus 4.6 (Adaptive)51.5%estimated ± 4.0 pp, high confidence
21Claude Mythos 5.251.2%estimated ± 2.7 pp, medium confidence
22GPT-5.451.1%estimated ± 4.0 pp, high confidence
23Claude Opus 4.850.9%estimated ± 4.0 pp, high confidence
24GPT-5.4 Pro50.5%estimated ± 5.6 pp, medium confidence
25Claude Sonnet 5.550.4%measured
26Gemini 3.1 Pro50.4%measured
27Claude Fable 5.150.4%estimated ± 4.0 pp, high confidence
28Grok 4.20 Multi-agent50.2%estimated ± 4.0 pp, high confidence
29Gemini 3 Flash49.7%measured
30Grok 4.1 Fast (Reasoning)49.0%estimated ± 4.0 pp, high confidence
31Kimi K2.648.9%estimated ± 4.0 pp, high confidence
32GPT-6.1 Terra48.8%measured
33Grok 4.2048.0%estimated ± 4.0 pp, high confidence
34Muse Spark 1.347.8%estimated ± 4.0 pp, high confidence
35Qwen3.7 Max47.8%estimated ± 4.0 pp, high confidence
36Muse Spark 1.147.7%estimated ± 4.0 pp, high confidence
37Claude Mythos 547.2%estimated ± 2.7 pp, medium confidence
38GPT-5.247.0%estimated ± 4.0 pp, high confidence
39Hy447.0%estimated ± 4.0 pp, high confidence
40GPT-5.6 Sol46.4%measured
41Kimi K3.1 Pro46.4%estimated ± 2.7 pp, medium confidence
42GPT-5.4 nano45.4%estimated ± 5.6 pp, medium confidence
43Grok 4.545.1%estimated ± 4.0 pp, high confidence
44Claude Mythos Preview45.0%estimated ± 2.7 pp, medium confidence
45Muse Spark 1.244.7%estimated ± 4.0 pp, high confidence
46Muse Spark Open44.3%estimated ± 4.0 pp, high confidence
47GPT-5.144.3%estimated ± 5.6 pp, medium confidence
48GLM-5.143.9%estimated ± 4.0 pp, high confidence
49GPT-6 Sol43.9%measured
50Claude Opus 4.743.9%estimated ± 5.6 pp, medium confidence
51Step 5 Preview43.3%estimated ± 4.0 pp, high confidence
52DeepSeek V4.1 Flash 101742.6%estimated ± 3.0 pp, high confidence
53Claude Sonnet 542.5%estimated ± 4.0 pp, high confidence
54Qwen3.7 Plus42.4%estimated ± 4.0 pp, high confidence
55Gemma 4 31B42.3%measured
56GLM-5.242.1%estimated ± 4.0 pp, high confidence
57GPT-542.1%estimated ± 5.6 pp, medium confidence
58o342.1%estimated ± 5.6 pp, medium confidence
59Qwen 3.6 Max (preview)42.0%estimated ± 4.0 pp, high confidence
60DeepSeek V4 Pro41.7%estimated ± 4.7 pp, high confidence
61Ling 3.1 Flash41.6%estimated ± 4.7 pp, high confidence
62GLM-5.341.5%estimated ± 2.7 pp, medium confidence
63Qwen3.5 397B41.5%estimated ± 5.8 pp, low confidence
64Grok 4.740.9%measured
65DeepSeek V4 Flash 073140.6%estimated ± 5.6 pp, medium confidence
66MiMo-V2.6-Pro40.0%estimated ± 2.7 pp, medium confidence
67GPT-5.6 Terra39.9%measured
68Qwen3.6-27B39.8%estimated ± 4.7 pp, high confidence
69Beam 1.0 501B-A23B39.7%estimated ± 4.7 pp, high confidence
70GPT-5.539.5%estimated ± 3.0 pp, high confidence
71Gemma 4 26B A4B39.5%measured
72Inkling-Small39.5%estimated ± 4.7 pp, high confidence
73Grok 4.639.3%measured
74Nemotron 3 Ultra39.3%estimated ± 4.7 pp, medium confidence
75MAI-Thinking-139.3%estimated ± 4.7 pp, medium confidence
76Hy4 preview38.9%estimated ± 4.0 pp, high confidence
77Claude Haiku 5.538.9%measured
78GPT-6 Luna38.7%measured
79Qwen3 Max38.6%estimated ± 5.8 pp, low confidence
80Seed 2.1 Turbo38.4%estimated ± 5.6 pp, medium confidence
81GPT-OSS 120B37.5%estimated ± 5.8 pp, low confidence
82GPT-5.6 Luna37.4%measured
83MiniMax M337.3%estimated ± 4.0 pp, high confidence
84SWE-2 Flash36.9%estimated ± 2.7 pp, medium confidence
85GLM-4.736.6%estimated ± 5.6 pp, medium confidence
86GLM-5.3-Flash36.6%estimated ± 5.6 pp, medium confidence
87Inkling36.6%estimated ± 5.6 pp, medium confidence
88Seed 2.1 Pro36.6%estimated ± 5.6 pp, medium confidence
89Kimi K336.2%estimated ± 2.7 pp, low confidence
90Gemini 3.5 Flash-Lite36.1%measured
91GPT-5.4 mini35.5%estimated ± 4.0 pp, high confidence
92Qwen3.6 Plus34.7%estimated ± 4.0 pp, medium confidence
93MiMo-V2.6-Flash34.6%estimated ± 2.7 pp, low confidence
94o3-mini34.4%estimated ± 5.6 pp, low confidence
95Kimi K2.7 Code34.0%estimated ± 5.6 pp, low confidence
96Qwen3.5 Flash34.0%estimated ± 5.6 pp, low confidence
97o4-mini33.6%estimated ± 5.6 pp, low confidence
98GPT-4.132.9%estimated ± 5.6 pp, low confidence
99o132.5%estimated ± 5.6 pp, low confidence
100GPT-4o32.2%estimated ± 5.6 pp, low confidence
101Qwen3.5 Plus32.2%estimated ± 5.6 pp, low confidence
102Grok 4.332.1%estimated ± 4.0 pp, medium confidence
103Claude Haiku 4.531.8%estimated ± 5.6 pp, low confidence
104Qwen3.8-27B31.7%estimated ± 4.0 pp, medium confidence
105Claude Opus 4.5 Thinking30.6%estimated ± 4.0 pp, medium confidence
106Claude 3 Opus30.3%estimated ± 5.6 pp, low confidence
107Qwen3.5-122B-A10B30.2%estimated ± 4.0 pp, medium confidence
108GPT-4o mini30.0%estimated ± 5.6 pp, low confidence
109GPT-4.1 mini29.2%estimated ± 5.6 pp, low confidence
110Claude Opus 4.529.0%estimated ± 4.0 pp, medium confidence
111GPT-4.1 nano28.9%estimated ± 5.6 pp, low confidence
112Qwen3.5-27B28.2%estimated ± 4.0 pp, medium confidence
113Claude Sonnet 4.626.6%estimated ± 4.0 pp, medium confidence
114Qwen3.7 Flash26.1%estimated ± 4.0 pp, medium confidence
115Claude Opus 4.625.8%estimated ± 4.0 pp, medium confidence
116Qwen3.6-35B-A3B24.9%estimated ± 4.0 pp, medium confidence
117Hy324.7%estimated ± 4.0 pp, medium confidence
118Step 3.7 Flash23.9%estimated ± 4.0 pp, medium confidence
119Claude Sonnet 4.5 Thinking22.7%estimated ± 4.0 pp, medium confidence
120Claude Sonnet 4.521.9%estimated ± 4.0 pp, medium confidence
121MiMo-V2.5-Pro21.1%estimated ± 4.0 pp, medium confidence
122MiniMax M3.1 Flash17.9%estimated ± 4.0 pp, medium confidence
123Mistral Large 417.4%estimated ± 4.0 pp, medium confidence
124MiMo-V2-Pro16.6%estimated ± 4.0 pp, medium confidence
125MiniMax M2.716.0%estimated ± 4.0 pp, medium confidence

All leaderboards

Agentic · terminal

Agentic · tools

Coding

Math

Knowledge & reasoning

Instruction following

Multilingual

Vision & documents

Long context

General