OnlySOTA

Arena — Vision

Shown as Arena publishes it — their metrics, their models. A model name opens its page here; the arrow opens the original board.

Published by
Arena
Licence
not stated
Updates
continuous
Data as of
2026-10-02
Fetched
2026-10-02
Entries
159 · 11 unmatched
Top 12 by Overall (style control)
  1. Claude Fable 5 1309
  2. Qwen 3.8 Max 1301
  3. Claude Opus 4.6 1299
  4. Claude Opus 4.7 1298
  5. Gemini 3.7 Flash 1296
  6. Muse Spark 1294
  7. Muse Spark 1.2 1293
  8. GPT 6.1 Sol 1291
  9. Gemini 3.8 Flash 1290
  10. Muse Spark 1.3 1290
  11. Gemini 3 Pro 1289
  12. Claude Opus 5 1289

An Elo score has no true zero, so instead of bars from 0 each row is a dot on a scale starting at 1287. Read the distance between dots, not the distance from the left edge.

159 scored

#Model
1Claude Fable 5highAnthropic13091301–1316
2Qwen 3.8 MaxAlibaba13011294–1309
3Claude Opus 4.6highAnthropic12991293–1306
4Claude Opus 4.7Anthropic12981292–1305
5Claude Opus 4.7highAnthropic12981291–1304
6Gemini 3.7 FlashhighGoogle12961286–1306
7Claude Opus 4.6Anthropic12941288–1301
8Muse SparkMeta12941284–1303
9Muse Spark 1.2xhighMeta12931279–1307
10GPT 6.1 SolmaxOpenAI12911275–1307
11Gemini 3.8 FlashhighGoogle12901279–1302
12Muse Spark 1.3maxMeta12901280–1301
13Gemini 3 ProGoogle12891282–1297
14Claude Opus 5highAnthropic12891282–1296
15Claude Fable 5.1maxAnthropic12881278–1298
16GPT 5.5OpenAI12871281–1292
17Claude Opus 4.8highAnthropic12861280–1293
18GPT 5.6 SolxhighOpenAI12851277–1292
19Gemini 3.5 FlashmediumGoogle12841277–1291
20GPT 6 AstramaxOpenAI12841273–1296
21Gemini 3.5 FlashhighGoogle12841277–1291
22GPT 5.4highOpenAI12841278–1290
23Muse Spark 1.1Meta12811274–1288
24GPT 5.5highOpenAI12811275–1287
25Gemini 3.6 FlashhighGoogle12801271–1288
26Claude Opus 4.8Anthropic12791273–1286
27Gemini 3.1 PropreviewGoogle12791274–1285
28Grok 4.5SpaceXAI12791272–1286
29GPT 5.4OpenAI12791272–1285
30GLM 5.3 FlashZ AI12781269–1288
31GPT 5.220260210 · chatOpenAI12781271–1285
32GPT 5.5 InstantOpenAI12761267–1285
33Claude Sonnet 4.6Anthropic12751269–1281
34Gemini 3 FlashGoogle12731268–1278
35Claude Sonnet 5.5xhighAnthropic12681253–1284
36GPT 5.6 TerraxhighOpenAI12681261–1276
37Qwen 3.7 PlusAlibaba12651259–1272
38Kimi K2.6Kimi12651258–1272
39Claude Sonnet 5highAnthropic12641257–1271
40Gemini 3.5 Flash LiteGoogle12641256–1273
41Grok 4.6highSpaceXAI12631254–1273
42Gemma 431BGoogle12621256–1268
43Gemini 3 Flashon · minimalGoogle12601254–1265
44GPT 5.6 LunaxhighOpenAI12601252–1267
45Step 5previewStepFun12601235–1284
46dola seed 2.0 proByteDance12571249–1264
47Grok 4.20beta · 0309 · onSpaceXAI12551249–1262
48Kimi K2.5onKimi12521246–1258
49GPT 5.1highOpenAI12511244–1259
50GPT 5.4 minihighOpenAI12511244–1257
51Grok 4.20beta · 0309 · multi-agentSpaceXAI12501244–1256
52Gemini 2.5 ProGoogle12471243–1252
53MiMo V2.6 FlashXiaomi12471234–1261
54Qwen 3.5397B A17BAlibaba12471242–1252
55MiMo V2.6 ProXiaomi12451231–1260
56Grok 4.3SpaceXAI12431237–1249
57GPT 5.2highOpenAI12431237–1249
58Qwen 3.827BAlibaba12431234–1251
59Gemma 426B A4BGoogle12421235–1248
60GPT 5.1OpenAI12401232–1247
61GPT 4o (ChatGPT)20250326OpenAI12391233–1244
62Kimi K2.5Kimi12371226–1248
63MiniMax M3MiniMax12371230–1243
64Gemini 3.1 Flash LiteGoogle12351229–1240
65MiMo V2.5Xiaomi12331227–1239
66Gemini 2.5 Flashpreview · 09-2025Google12301220–1240
67GLM 5V TurboZ AI12301225–1236
68GPT 5.2OpenAI12291223–1235
69Qwen 3.5122B A10BAlibaba12261220–1233
70GPT 5OpenAI12261219–1233
71GPT 4.5preview · 2025-02-27OpenAI12251213–1237
72ERNIE 5.0preview · 1220Baidu12191208–1230
73Qwen 3.527BAlibaba12161211–1222
74MiMo V2 OmniXiaomi12161209–1223
75o32025-04-16OpenAI12161209–1222
76Gemini 2.5 FlashGoogle12151210–1220
77Qwen 3 VL235B A22BAlibaba12151208–1222
78GPT 4.12025-04-14OpenAI12141207–1221
79GPT 5highOpenAI12101203–1217
80Claude Opus 420250514 · on · 16000Anthropic12071191–1223
81Inkling SmallThinking Machines12061198–1215
82Claude Sonnet 420250514 · on · 32000Anthropic12061189–1223
83o4 mini2025-04-16OpenAI12011194–1208
84GPT 4.1 mini2025-04-14OpenAI12011193–1208
85Mistral Large 3Mistral12011192–1209
86Mistral Medium 3.5Mistral11991190–1209
87GPT 5.4 nanohighOpenAI11991193–1206
88Grok 4.1 FastonSpaceXAI11961189–1203
89Claude Sonnet 3.720250219 · on · 32000Anthropic11931177–1208
90o12024-12-17OpenAI11931182–1204
91Claude Opus 420250514Anthropic11881176–1201
92Gemini 2.5 Flash Litepreview · 06-17 · onGoogle11881181–1196
93Claude Sonnet 420250514Anthropic11881174–1202
94Qwen 3 VL235B A22B · onAlibaba11881176–1200
95Qwen VL Max2025-08-13Alibaba11851173–1197
96Grok 40709SpaceXAI11841177–1192
97GPT 5 minihighOpenAI11831175–1191
98Gemini 1.5 ProGoogle11791170–1188
99Claude Sonnet 3.720250219Anthropic11751166–1185
100Gemini 2.5 Flash Litepreview · 09-2025 · offGoogle11741165–1184

What each column measures

Overall (style control)
Higher is better · Multimodal