OnlySOTA

Arena — Vision

As Arena publishes it — their metrics, their models. Follow a model name to its page upstream.

Published by
Arena
Licence
not stated
Updates
continuous
Data as of
2026-08-21
Fetched
2026-08-22
Entries
147 · 72 unmatched
Top 12 by Overall (style control), one row per model
  1. Claude Fable 5 (Opus 4.8 fallback) 1312
  2. qwen3.8-max Unknown developerUD 1302
  3. Claude Opus 4.7 1301
  4. Claude Opus 4.6 1299
  5. Muse Spark 1294
  6. Muse Spark 1.2 1292
  7. Claude Opus 5 1292
  8. Gemini 3 Pro Preview 1289
  9. Gemini 3.5 Flash 1286
  10. Claude Opus 4.8 1285
  11. Gemini 3.6 Flash 1285
  12. GPT-5.5 1285

An Elo has no zero — it is an arbitrary anchor, not "no skill" — so a bar drawn from it would encode the anchor. Each row is a dot on a range beginning at 1282; the distance between rows is the reading, the distance from the left edge is not.

147 scored

#Model
1Claude Fable 5 (Opus 4.8 fallback)anthropic13121303–1320
2Unknown developerUDqwen3.8-maxmaxUnmatched13021293–1310
3Claude Opus 4.7highanthropic13011294–1308
4Claude Opus 4.7anthropic12991292–1306
5Claude Opus 4.6highanthropic12991292–1305
6Muse Sparkmeta12941285–1303
7Claude Opus 4.6anthropic12931287–1300
8Muse Spark 1.2xhighmeta12921277–1307
9Claude Opus 5highanthropic12921283–1301
10Gemini 3 Pro Previewgoogle12891282–1297
11Gemini 3.5 Flashhighgoogle12861278–1295
12Claude Opus 4.8highanthropic12851277–1293
13Gemini 3.6 Flashhighgoogle12851273–1297
14GPT-5.5openai12851278–1292
15Gemini 3.5 Flashmediumgoogle12851276–1293
16GPT-5.5highopenai12821276–1289
17Grok 4.5xai12821273–1291
18Muse Spark 1.1meta12811272–1290
19GPT-5.6 Solxhighopenai12811271–1291
20GPT-5.4openai12801273–1287
21GPT-5.4highopenai12801273–1286
22Claude Opus 4.8anthropic12781271–1286
23Unknown developerUDgpt-5.2-chat-latest-20260210Unmatched12781271–1285
24Gemini 3.1 Pro Previewgoogle12771271–1283
25Unknown developerUDgpt-5.5-instantUnmatched12771268–1286
26Claude Sonnet 4.6anthropic12751268–1281
27Claude Sonnet 5highanthropic12711262–1279
28Gemini 3 Flash Previewgoogle12711265–1276
29Gemini 3.5 Flash-Litegoogle12701258–1282
30GPT-5.6 Terraxhighopenai12661256–1275
31Qwen3.7 Plusalibaba12651257–1274
32Kimi K2.6kimi12631256–1270
33Gemma 4 31Bgoogle12601254–1267
34Gemini 3 Flash Previewon · minimalgoogle12601254–1265
35Unknown developerUDdola-seed-2.0-proUnmatched12581250–1265
36Unknown developerUDgrok-4.20-beta-0309-reasoningonUnmatched12561250–1263
37Unknown developerUDqwen3.8-27bUnmatched12531240–1266
38GPT-5.6 Lunaxhighopenai12531243–1262
39GPT-5.4 minihighopenai12521245–1258
40Unknown developerUDgrok-4.20-multi-agent-beta-0309Unmatched12521245–1258
41GPT-5.1highopenai12501242–1259
42Kimi K2.5onkimi12501244–1256
43Qwen3.5 397B A17Balibaba12481242–1254
44Gemini 2.5 Progoogle12461241–1251
45GPT-5.2highopenai12441238–1251
46Gemma 4 26B A4Bgoogle12421235–1248
47Grok 4.3xai12411234–1248
48GPT-4o (March 2025, chatgpt-4o-latest)openai12411235–1246
49GPT-5.1openai12381230–1246
50Unknown developerUDkimi-k2.5-instantUnmatched12381227–1249
51MiniMax-M3minimax12371229–1245
52MiMo-V2.5xiaomi12351228–1242
53Gemini 3.1 Flash-Litegoogle12341228–1241
54GLM 5V Turbozai12301224–1236
55GPT-5.2openai12301223–1236
56Qwen3.5 122B A10Balibaba12271220–1234
57Unknown developerUDgpt-5-chatUnmatched12261219–1233
58Unknown developerUDgpt-4.5-preview-2025-02-27Unmatched12261214–1237
59Gemini 2.5 Flash Preview (Sep '25)google12251215–1235
60Qwen3.5 27Balibaba12191213–1225
61Unknown developerUDernie-5.0-preview-1220Unmatched12191207–1230
62MiMo-V2-Omnixiaomi12171210–1225
63Unknown developerUDo3-2025-04-16Unmatched12171210–1224
64Qwen3 VL 235B A22B Instructalibaba12151208–1222
65Gemini 2.5 Flashgoogle12141210–1219
66Unknown developerUDgpt-4.1-2025-04-14Unmatched12141207–1221
67GPT-5highopenai12121205–1220
68Unknown developerUDclaude-sonnet-4-20250514-thinking-32kUnmatched12081191–1225
69Unknown developerUDclaude-opus-4-20250514-thinking-16kUnmatched12071191–1223
70Inkling Smallthinking-machines12061194–1218
71Mistral Large 3mistral12031192–1213
72Unknown developerUDgpt-4.1-mini-2025-04-14Unmatched12021195–1210
73Unknown developerUDo4-mini-2025-04-16Unmatched12011194–1209
74GPT-5.4 nanohighopenai12011195–1208
75Mistral Medium 3.5mistral11991189–1209
76Grok 4.1 Fastonxai11951188–1203
77Unknown developerUDclaude-3-7-sonnet-20250219-thinking-32kUnmatched11951180–1211
78Unknown developerUDo1-2024-12-17Unmatched11931182–1204
79Qwen3 VL 235B A22Bonalibaba11891177–1202
80Unknown developerUDclaude-sonnet-4-20250514Unmatched11891175–1203
81Unknown developerUDclaude-opus-4-20250514Unmatched11891176–1202
82Unknown developerUDgemini-2.5-flash-lite-preview-06-17-thinkingonUnmatched11881181–1196
83Unknown developerUDqwen-vl-max-2025-08-13Unmatched11871175–1199
84GPT-5 minihighopenai11831174–1191
85Unknown developerUDgrok-4-0709Unmatched11821175–1189
86Unknown developerUDgemini-1.5-pro-002Unmatched11791170–1188
87Unknown developerUDclaude-3-7-sonnet-20250219Unmatched11751166–1185
88Gemini 2.5 Flash-Lite Preview (Sep '25)offgoogle11751165–1184
89Unknown developerUDgemini-2.0-flash-001Unmatched11721165–1179
90GLM-4.6Vzai11641150–1178
91GPT-4o (May '24)openai11621153–1171
92Unknown developerUDclaude-3-5-sonnet-20241022Unmatched11611153–1169
93Gemma 3 27B Instructgoogle11601151–1168
94Unknown developerUDhunyuan-vision-1.5-thinkingonUnmatched11591147–1171
95Unknown developerUDmistral-medium-2508Unmatched11591152–1165
96Unknown developerUDstep-1o-turbo-202506Unmatched11571143–1171
97Unknown developerUDmistral-medium-2505Unmatched11561148–1164
98GLM-4.5Vzai11551144–1167
99Unknown developerUDhunyuan-large-visionUnmatched11511135–1168
100GPT-5 nanohighopenai11481137–1159
101Llama 4 Maverickmeta11471137–1156
102Unknown developerUDstep-3Unmatched11451133–1157
103Claude 3.5 Sonnet (June '24)anthropic11451135–1155
104Unknown developerUDgemini-1.5-flash-002Unmatched11411131–1151
105Mistral Small 3.2mistral11411131–1150
106Unknown developerUDgemini-2.0-flash-lite-preview-02-05Unmatched11361126–1147
107Unknown developerUDclaude-3-5-haiku-20241022Unmatched11281112–1144
108Unknown developerUDmistral-small-3.1-24b-instruct-2503Unmatched11281119–1136
109Llama 4 Scoutmeta11281118–1137
110Unknown developerUDstep-1o-vision-32k-highresUnmatched11251112–1138
111Unknown developerUDqwen2.5-vl-72b-instructUnmatched11221111–1132
112Unknown developerUDgemini-1.5-pro-001Unmatched11201108–1132
113Unknown developerUDqwen2.5-vl-32b-instructUnmatched11201104–1135
114GPT-4o (Aug '24)openai11191107–1131
115Unknown developerUDgpt-4-turbo-2024-04-09Unmatched11121101–1124
116Molmo2-8Bai211091089–1129
117Unknown developerUDgpt-4o-mini-2024-07-18Unmatched10981090–1106
118Pixtral Largemistral10951086–1105
119Unknown developerUDgpt-4.1-nano-2025-04-14Unmatched10891071–1108
120Unknown developerUDqwen-vl-max-1119Unmatched10851069–1101
121Unknown developerUDqwen2-vl-72bUnmatched10851075–1095
122Unknown developerUDgemini-1.5-flash-8b-001Unmatched10721061–1082
123Unknown developerUDstep-1v-32kUnmatched10641048–1081
124Unknown developerUDclaude-3-opus-20240229Unmatched10621052–1072
125Unknown developerUDgemini-1.5-flash-001Unmatched10601048–1072
126Unknown developerUDmolmo-72b-0924Unmatched10451032–1058
127Unknown developerUDhunyuan-standard-vision-2024-12-31Unmatched10421021–1063
128Unknown developerUDllama-3.2-vision-90b-instructUnmatched10331023–1042
129Unknown developerUDqwen2-vl-7b-instructUnmatched10311021–1042
130Unknown developerUDpixtral-12b-2409Unmatched10251016–1035
131Unknown developerUDinternvl2-26bUnmatched10251012–1038
132Unknown developerUDamazon-nova-pro-v1.0Unmatched10191005–1033
133Unknown developerUDamazon-nova-lite-v1.0Unmatched10191004–1034
134Unknown developerUDclaude-3-sonnet-20240229Unmatched10171005–1028
135Unknown developerUDyi-visionUnmatched1002984–1021
136Unknown developerUDclaude-3-haiku-20240307Unmatched1000988–1013
137Unknown developerUDc4ai-aya-vision-32bUnmatched1000978–1022
138Molmo 7B-Dai2994980–1008
139Unknown developerUDllama-3.2-vision-11b-instructUnmatched993982–1004
140Unknown developerUDnvila-internal-15b-v1Unmatched986966–1006
141Unknown developerUDllava-onevision-qwen2-72b-ovUnmatched979961–997
142Unknown developerUDllava-v1.6-34bUnmatched964952–977
143Unknown developerUDcogvlm2-llama3-chat-19bUnmatched964949–979
144Unknown developerUDminicpm-v-2_6Unmatched963947–979
145Unknown developerUDinternvl2-4bUnmatched958946–971
146Unknown developerUDphi-3.5-vision-instructUnmatched920904–936
147Unknown developerUDphi-3-vision-128k-instructUnmatched882863–900

What the columns measure

Overall (style control)
Higher is better · Multimodal