OnlySOTA

Arena — Code

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
118 · 23 unmatched
Top 12 by Overall, one row per model
  1. Claude Opus 5 1691
  2. Kimi K3 1674
  3. qwen3.8-max Unknown developerUD 1669
  4. Grok 4.6 1629
  5. Claude Fable 5 (Opus 4.8 fallback) 1626
  6. GPT-5.6 Sol 1619
  7. GLM-5.3 1599
  8. qwen3.8-27b Unknown developerUD 1595
  9. Gemini 3.7 Flash 1587
  10. GLM-5.2 1582
  11. deepseek-v4-pro-high-20260813 Unknown developerUD 1582
  12. deepseek-v4-flash-high Unknown developerUD 1579

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 1566; the distance between rows is the reading, the distance from the left edge is not.

118 scored

#Model
1Claude Opus 5maxanthropic16911682–1699
2Kimi K3maxkimi16741662–1685
3Unknown developerUDqwen3.8-maxmaxUnmatched16691656–1681
4Claude Opus 5highanthropic16631655–1671
5Grok 4.6highxai16291612–1646
6Claude Fable 5 (Opus 4.8 fallback)anthropic16261618–1634
7GPT-5.6 Solcodex-harness · xhighopenai16191611–1626
8GLM-5.3maxzai15991583–1614
9Unknown developerUDqwen3.8-27bUnmatched15951582–1608
10Gemini 3.7 Flashhighgoogle15871574–1600
11GLM-5.2maxzai15821574–1590
12Unknown developerUDdeepseek-v4-pro-high-20260813Unmatched15821570–1594
13Unknown developerUDdeepseek-v4-flash-highhighUnmatched15791568–1590
14Claude Opus 4.8highanthropic15631556–1571
15Claude Opus 4.7anthropic15581551–1564
16Claude Opus 4.7highanthropic15571550–1563
17Grok 4.5xai15561548–1565
18Claude Opus 4.6highanthropic15461540–1551
19Claude Opus 4.8anthropic15391532–1546
20Muse Spark 1.1meta15391531–1548
21Gemini 3.6 Flashhighgoogle15391530–1547
22Claude Sonnet 5highanthropic15391531–1547
23Claude Opus 4.6anthropic15361531–1542
24Muse Spark 1.2xhighmeta15341520–1548
25Claude Sonnet 4.6anthropic15221517–1528
26Unknown developerUDseed-2.1-pro-previewUnmatched15211514–1529
27GPT-5.6 Terracodex-harness · xhighopenai15201511–1529
28Hy3tencent15181506–1531
29GPT-5.6 Lunacodex-harness · xhighopenai15181509–1526
30Unknown developerUDqwen3.7-max-20260517Unmatched15171509–1524
31GLM-5.1zai15091502–1516
32Kimi K2.6kimi15091501–1516
33GPT-5.5codex-harness · xhighopenai15081502–1515
34Gemini 3.5 Flashhighgoogle14991491–1507
35Unknown developerUDclaude-opus-4-5-20251101-high-32kUnmatched14941486–1503
36Gemini 3.5 Flashmediumgoogle14901483–1497
37MiniMax-M3minimax14881481–1495
38GPT-5.5codex-harness · highopenai14861480–1492
39Unknown developerUDqwen3.6-max-previewUnmatched14791466–1492
40MiMo-V2.5-Proxiaomi14761469–1482
41Kimi K2.7 Codekimi14731463–1482
42Unknown developerUDclaude-opus-4-5-20251101Unmatched14681461–1475
43Unknown developerUDdeepseek-v4-pro-high-previewUnmatched14641457–1470
44GPT-5.4codex-harness · highopenai14631443–1482
45Qwen3.6 Plusalibaba14601454–1466
46GPT-5.5codex-harnessopenai14581451–1464
47Gemini 3.5 Flash-Litegoogle14491407–1492
48Gemini 3.1 Pro Previewgoogle14461441–1452
49DeepSeek V4 Pro 0813deepseek14451439–1452
50GPT-5.4codex-harness · mediumopenai14421423–1461
51Gemini 3 Flash Previewgoogle14381429–1447
52Gemini 3 Pro Previewgoogle14381429–1446
53MiMo-V2.5xiaomi14381431–1444
54Kimi K2.5onkimi14361431–1441
55GLM-5zai14351427–1444
56GLM-4.7zai14341422–1447
57MiMo-V2-Proxiaomi14341425–1442
58Unknown developerUDdeepseek-v4-flash-high-previewUnmatched14311424–1438
59GPT-5mediumopenai14191403–1435
60GPT-5.2openai14181395–1441
61GPT-5.3 Codexcodex-harnessopenai14081395–1422
62Inklingthinking-machines14081400–1416
63Unknown developerUDkimi-k2.5-instantUnmatched14051393–1417
64Inkling Smallthinking-machines14021391–1412
65GLM 5V Turbozai14001386–1414
66Qwen3.5 397B A17Balibaba13991394–1405
67MiniMax-M2.7minimax13981392–1404
68GPT-5.4 minihighopenai13971390–1404
69Unknown developerUDclaude-sonnet-4-5-20250929-high-32kUnmatched13921385–1400
70GPT-5.1mediumopenai13911379–1403
71GPT-5.4openai13911376–1405
72Unknown developerUDclaude-opus-4-1-20250805Unmatched13891378–1400
73Unknown developerUDminimax-m2.1-previewUnmatched13871377–1397
74Unknown developerUDclaude-sonnet-4-5-20250929Unmatched13861378–1393
75MiniMax-M2.5minimax13841375–1392
76Gemini 3 Flash Previewon · minimalgoogle13831377–1388
77Unknown developerUDgrok-4.20-beta-0309-reasoningonUnmatched13741367–1380
78Solar Pro 4upstage13711354–1388
79GPT-5.3 Codexcodex-harnessopenai13691357–1381
80Gemma 4 31Bgoogle13641356–1371
81Gemma 4 26B A4Bgoogle13621344–1379
82DeepSeek V3.2ondeepseek13601351–1370
83Muse Glimmermeta13591343–1375
84Qwen3.5 122B A10Balibaba13581350–1365
85Qwen3.5 27Balibaba13571349–1366
86Grok 4.3xai13571350–1363
87Hy3-previewtencent13561339–1374
88Unknown developerUDlaguna-m.1Unmatched13471338–1357
89GPT-5.1openai13411332–1349
90GLM-4.6zai13401329–1351
91GPT-5.2 Codexopenai13381329–1347
92GPT-5.1 Codexopenai13361325–1348
93MiMo-V2-Flashoffxiaomi13301320–1340
94Unknown developerUDclaude-haiku-4-5-20251001Unmatched13271322–1332
95DeepSeek V3.2deepseek13251316–1333
96Kimi K2 Thinkingkimi13231315–1330
97Unknown developerUDlaguna-xs.2Unmatched13021291–1313
98MiniMax-M2minimax12971286–1308
99MiMo-V2-Flashonxiaomi12921276–1309
100Qwen3 Coder 480B A35B Instructalibaba12731265–1281
101DeepSeek V3.2 Expdeepseek12721258–1285
102Mistral Medium 3.5mistral12651250–1280
103KAT-Coder-Pro V1kwaikat12551236–1275
104Gemini 3.1 Flash-Litegoogle12541247–1261
105Qwen3.5 35B A3Balibaba12501232–1268
106GPT-5.1 Codex miniopenai12441221–1266
107Grok 4.1 Fastonxai12401229–1251
108Trinity Large Thinkingarcee12381218–1259
109Unknown developerUDqwen3.5-flashUnmatched12381218–1258
110Mistral Large 3mistral12301204–1256
111Gemini 2.5 Progoogle12261210–1242
112Unknown developerUDgrok-4.1-thinkingonUnmatched12101185–1235
113Devstral 2mistral11941172–1215
114Granite 4.1 8Bibm11911172–1210
115Mercury 2inception11661141–1191
116Grok Code Fast 1xai11651136–1193
117Grok 4 Fastonxai11621134–1190
118Unknown developerUDdevstral-medium-2507Unmatched10801049–1111

What the columns measure

Overall
Higher is better · Coding