OnlySOTA

Arena — Code

按 Arena 发布的原样呈现,指标和模型都以它为准。

发布方
Arena
许可协议
未声明
更新频率
持续更新
数据截至
2026-10-08
本站抓取时间
2026-10-09
条目数
142
按 Overall 排名前 12 的模型
  1. Claude Opus 5.5 1813
  2. GPT 6 Astra 1786
  3. Claude Sonnet 5.5 1774
  4. GPT 6.1 Sol 1755
  5. Claude Fable 5.1 1744
  6. Claude Opus 5 1691
  7. GPT 6 Sol 1688
  8. Gemini 4 Argon 1678
  9. Qwen 3.8 Max 1674
  10. Muse Spark 1.3 1657
  11. Kimi K3 1654
  12. Grok 4.7 1639

Elo 分没有零点,所以每行是从 1618 起的刻度上的一个点:看点与点之间的距离,不看离左边多远。

142 个有分数

#模型
1Claude Opus 5.5maxAnthropic18131798–1828
2GPT 6 AstramaxOpenAI17861775–1797
3Claude Sonnet 5.5xhighAnthropic17741759–1789
4GPT 6.1 SolmaxOpenAI17551740–1770
5Claude Fable 5.1maxAnthropic17441734–1753
6Claude Sonnet 5.5highAnthropic17151704–1726
7Claude Opus 5maxAnthropic16911685–1698
8GPT 6 SolmaxOpenAI16881676–1700
9Gemini 4 ArgonhighGoogle16781662–1693
10Qwen 3.8 Max0902Alibaba16741666–1681
11Qwen 3.8 MaxAlibaba16721660–1684
12Claude Opus 5highAnthropic16571651–1662
13Muse Spark 1.3maxMeta16571649–1665
14Kimi K3maxKimi16541648–1661
15Grok 4.7xhighSpaceXAI16391628–1650
16MiMo V2.6 FlashXiaomi16371626–1647
17Qwen 3.8 Flash NextAlibaba16331624–1641
18Hy4previewTencent16321622–1641
19MiMo V2.6 ProXiaomi16291618–1640
20Muse Spark 1.3xhighMeta16261617–1634
21Claude Fable 5highAnthropic16251618–1631
22GLM 5.3maxZ AI16221614–1630
23DeepSeek V4.1 FlashmaxDeepSeek16191608–1630
24GPT 5.6 Solcodex-harness · xhighOpenAI16181612–1624
25Grok 4.6highSpaceXAI16171609–1625
26GLM 5.3 FlashZ AI16091602–1616
27GLM 5.2maxZ AI16031597–1609
28Qwen 3.827BAlibaba15931586–1600
29Gemini 3.7 FlashhighGoogle15921584–1599
30Claude Haiku 5.5highAnthropic15871567–1607
31Gemini 3.8 FlashhighGoogle15831576–1591
32DeepSeek V4 Pro20260813 · highDeepSeek15821573–1591
33DeepSeek V4 FlashhighDeepSeek15821572–1591
34GPT 6 LunamaxOpenAI15811574–1588
35Step 5preview · highStepFun15641552–1575
36Claude Opus 4.7highAnthropic15571552–1563
37Claude Opus 4.8highAnthropic15561550–1562
38Claude Opus 4.7Anthropic15551550–1561
39Grok 4.5SpaceXAI15531546–1560
40Claude Opus 4.6highAnthropic15471541–1552
41Muse Spark 1.1Meta15421535–1550
42Mistral Large 4Mistral15411525–1558
43Claude Sonnet 5highAnthropic15411535–1547
44Gemini 3.6 FlashhighGoogle15381530–1545
45Claude Opus 4.6Anthropic15371532–1543
46Claude Opus 4.8Anthropic15361530–1541
47Muse Spark 1.2xhighMeta15331519–1546
48Claude Sonnet 4.6Anthropic15221517–1528
49GPT 5.6 Terracodex-harness · xhighOpenAI15221515–1528
50Seed 2.1 PropreviewByteDance15191513–1525
51GPT 5.6 Lunacodex-harness · xhighOpenAI15191513–1525
52Qwen 3.7 Max20260517Alibaba15151508–1522
53GPT 5.5codex-harness · xhighOpenAI15131507–1519
54Gemini 3.5 FlashGoogle15111491–1530
55Kimi K2.6Kimi15091502–1516
56Hunyuan Hy3Tencent15081500–1517
57GLM 5.1Z AI15081502–1515
58Gemini 3.5 FlashhighGoogle14991492–1506
59Claude Opus 4.520251101 · high · 32000Anthropic14941486–1502
60Gemini 3.5 FlashmediumGoogle14911484–1498
61GPT 5.5codex-harness · highOpenAI14871482–1493
62Qwen 3.6 MaxpreviewAlibaba14821470–1495
63MiniMax M3MiniMax14821476–1487
64MiMo V2.5 ProXiaomi14791474–1485
65Kimi K2.7 CodeKimi14731464–1483
66Claude Opus 4.520251101Anthropic14691462–1476
67GPT 5.4codex-harness · highOpenAI14651446–1484
68DeepSeek V4 Propreview · highDeepSeek14641457–1470
69Qwen 3.6 PlusAlibaba14611456–1467
70GPT 5.5codex-harnessOpenAI14561450–1462
71DeepSeek V4 ProDeepSeek14471440–1453
72Gemini 3.1 PropreviewGoogle14471442–1452
73GPT 5.4codex-harness · mediumOpenAI14441426–1463
74Gemini 3.5 Flash LiteGoogle14401399–1480
75Gemini 3 ProGoogle14401431–1448
76Gemini 3 FlashGoogle14391431–1448
77MiMo V2.5Xiaomi14381432–1444
78Kimi K2.5onKimi14361431–1442
79GLM 4.7Z AI14351423–1447
80GLM 5Z AI14341426–1443
81MiMo V2 ProXiaomi14331425–1441
82DeepSeek V4 Flashpreview · highDeepSeek14311424–1438
83GPT 5mediumOpenAI14181402–1433
84GPT 5.2OpenAI14161394–1438
85InklingThinking Machines14131407–1419
86Inkling SmallThinking Machines14091400–1418
87GPT 5.3 Codexcodex-harnessOpenAI14091395–1422
88Kimi K2.5Kimi14041392–1416
89GLM 5V TurboZ AI14011388–1414
90GPT 5.4OpenAI14001390–1411
91Qwen 3.5397B A17BAlibaba14001395–1405
92MiniMax M2.7MiniMax13981392–1404
93GPT 5.4 minihighOpenAI13971391–1404
94GPT 5.1mediumOpenAI13951384–1407
95Claude Sonnet 4.520250929 · high · 32000Anthropic13931386–1401
96Claude Opus 4.120250805Anthropic13901379–1401
97MiniMax M2.5MiniMax13871378–1395
98Gemini 3 Flashon · minimalGoogle13851379–1390
99Claude Sonnet 4.520250929Anthropic13851377–1392
100MiniMax M2.1previewMiniMax13841375–1394

每一列测的是什么

Overall
越高越好 · 编程