Overall Model Rankings

Updated 8 minutes ago

Average performance across all supported visiontasks.

Votes power rankings.

Top Model Scores

Overall ELO ratings averaged across all tasks

Loading chart...

Performance vs Accuracy

ELO score vs average latency • Better models are top-left

Loading chart...


1	SAM 3	multimodal	1402	1	2.68 s	Meta
2	Gemini 2.5 Pro	multimodal	1269	5	15.82 s	Google
3	Gemini 3 Flash	multimodal	1265	5	8.41 s	Google
4	Gemini 3.5 Flash	multimodal	1255	2	23.55 s	Google
5	Claude Fable 5	multimodal	1249	1	9.33 s	Anthropic
6	Gemini 2.5 Flash	multimodal	1248	5	8.04 s	Google
7	GPT-5.5	multimodal	1243	3	8.61 s	OpenAI
8	GPT-5.4	multimodal	1235	5	5.56 s	OpenAI
9	GPT-5 Mini	multimodal	1227	5	17.36 s	OpenAI
10	Gemini 3.1 Pro	multimodal	1226	4	16.50 s	Google
11	Qwen3.5 122B A10B	multimodal	1223	3	17.24 s	Qwen
12	Qwen VL Max	multimodal	1218	3	11.90 s	Qwen
13	Qwen3 VL 235B A22B Instruct	multimodal	1215	3	7.13 s	Qwen
14	Qwen3 VL 30B A3B Instruct	multimodal	1214	3	9.87 s	Qwen
15	Grok 4	multimodal	1213	3	25.29 s	xAI
16	GPT-5	multimodal	1212	5	21.52 s	OpenAI
17	GPT-5.2	multimodal	1212	5	5.96 s	OpenAI
18	GPT-5.6 Sol	multimodal	1212	2	9.58 s	OpenAI
19	Gemma 4 26B A4B	multimodal	1208	4	3.95 s	Google
20	Kimi K2.5	multimodal	1206	2	38.79 s	Moonshot AI
21	GPT-5.6 Luna	multimodal	1206	3	5.71 s	OpenAI
22	Claude Sonnet 5	multimodal	1206	3	5.95 s	Anthropic
23	Qwen3.5 9b	multimodal	1205	2	29.12 s	Qwen
24	GPT-5.4 Mini	multimodal	1205	4	2.69 s	OpenAI
25	GPT-5.1	multimodal	1204	5	6.68 s	OpenAI
26	Qwen3.6 Plus	multimodal	1204	3	19.82 s	Qwen
27	Qwen3.6 35B A3B	multimodal	1203	4	19.56 s	Qwen
28	Qwen3.5 27B	multimodal	1203	3	31.16 s	Qwen
29	Claude Opus 4.7	multimodal	1202	5	8.60 s	Anthropic
30	Gemma 4 31B	multimodal	1201	5	17.66 s	Google
31	Qwen3.6 27B	multimodal	1201	2	20.46 s	Qwen
32	Muse Spark 1.1	multimodal	1200	1	16.62 s	Meta
33	Claude Opus 4.8	multimodal	1200	2	9.12 s	Anthropic
34	GPT-5.6 Terra	multimodal	1199	1	6.82 s	OpenAI
35	Qwen3.5 397B A17B	multimodal	1197	3	20.60 s	Qwen
36	Qwen3 VL 8B Instruct	multimodal	1197	3	7.97 s	Qwen
37	Claude Opus 4.6	multimodal	1195	5	7.45 s	Anthropic
38	Gemini 3.1 Flash-Lite	multimodal	1192	2	3.31 s	Google
39	Gemma 3 27B	multimodal	1192	3	9.81 s	Google
40	Llama 4 Scout	multimodal	1190	3	4.17 s	Meta
41	Claude Opus 4.5	multimodal	1190	5	7.51 s	Anthropic
42	Claude Sonnet 4.5	multimodal	1190	5	9.14 s	Anthropic
43	Qwen3.5 35B A3B	multimodal	1190	3	19.51 s	Qwen
44	Gemini 2.5 Flash-Lite	multimodal	1189	5	2.27 s	Google
45	Claude Sonnet 4.6	multimodal	1188	5	7.11 s	Anthropic
46	Qwen3.6 Flash	multimodal	1188	1	32.69 s	Qwen
47	Mistral Medium 3.1	multimodal	1187	3	5.94 s	Mistral
48	Gemma 3 12B	multimodal	1187	3	14.55 s	Google
49	Mistral Small 3.1 24B	multimodal	1181	3	6.54 s	Mistral
50	Claude Opus 4.1	multimodal	1178	5	10.97 s	Anthropic
51	YOLO World	multimodal	1174	1	3.17 s	Tencent AI Lab
52	Pixtral 12B	multimodal	1172	3	8.56 s	Mistral
53	Llama 4 Maverick	multimodal	1172	3	4.18 s	Meta
54	GPT-5 Nano	multimodal	1172	5	16.42 s	OpenAI
55	Claude Haiku 4.5	multimodal	1167	5	3.88 s	Anthropic
56	GPT-5.4 Nano	multimodal	1163	3	3.68 s	OpenAI
57	Gemma 3 4B	multimodal	1161	3	10.27 s	Google
58	Llama 3.2 Vision 11b	multimodal	1160	4	9.78 s	Meta
59	Google Vision OCR	vision	1151	1	869 ms	Google
60	Qwen2.5 VL 7B Instruct	multimodal	1149	3	5.98 s	Qwen
61	Florence-2	multimodal	1110	3	6.49 s	Microsoft