Code Arena | Overall

View overall rankings across AI models on agentic coding tasks involving multi-step reasoning, tool use, and production-style workflows.

Mar 4, 2026

184,734 votes

51 models

	Rank Spread
1	12	claude-opus-4-6-thinking Anthropic · Proprietary	1556+14/-14	2,553
2	12	claude-opus-4-6 Anthropic · Proprietary	1555+13/-13	3,273
3	33	claude-sonnet-4-6 Anthropic · Proprietary	1523+12/-12	3,150
4	44	claude-opus-4-5-20251101-thinking-32k Anthropic · Proprietary	1497+8/-8	11,859
5	57	claude-opus-4-5-20251101 Anthropic · Proprietary	1475+7/-7	11,994
6	58	gpt-5.2-high OpenAI · Proprietary	1472+16/-16	1,696
7	512	gemini-3.1-pro-preview Google · Proprietary	1461+15/-15	1,829
8	612	glm-5 Z.ai · MIT	1447+12/-12	3,166
9	712	gemini-3-pro Google · Proprietary	1442+7/-7	17,443
10	712	gemini-3-flash Google · Proprietary	1442+7/-7	13,601
11	713	glm-4.7 Z.ai · MIT	1441+10/-10	5,131
12	713	kimi-k2.5-thinking Moonshot · Modified MIT	1438+10/-10	4,626
13	1114	minimax-m2.5 MiniMax · Modified MIT	1422+10/-10	4,259
14	1319	kimi-k2.5-instant Moonshot · Modified MIT	1415+11/-11	3,539
15	1423	gemini-3-flash (thinking-minimal) Google · Proprietary	1404+8/-8	9,579
16	1423	minimax-m2.1-preview MiniMax · MIT	1402+8/-8	9,793
17	1426	gpt-5.2 OpenAI · Proprietary	1396+15/-16	1,634
18	1425	qwen3.5-397b-a17b Alibaba · Apache 2.0	1396+11/-11	3,136
19	1426	gpt-5-medium OpenAI · Proprietary	1394+12/-12	3,929
20	1525	claude-sonnet-4-5-20250929-thinking-32k Anthropic · Proprietary	1390+7/-7	14,795
21	1526	claude-opus-4-1-20250805 Anthropic · Proprietary	1390+8/-8	8,983
22	1526	gpt-5.1-medium OpenAI · Proprietary	1389+9/-9	6,441
23	1726	claude-sonnet-4-5-20250929 Anthropic · Proprietary	1386+7/-7	16,570
24	1526	qwen3.5-122b-a10b Alibaba · Apache 2.0	1384+15/-15	1,675
25	1727	qwen3.5-27b Alibaba · Apache 2.0	1375+16/-16	1,593
26	1927	deepseek-v3.2-thinking DeepSeek · MIT	1373+9/-9	6,560
27	2530	glm-4.6 Z.ai · MIT	1357+8/-8	8,746
28	2732	gpt-5.1 OpenAI · Proprietary	1344+7/-7	13,268
29	2732	mimo-v2-flash (non-thinking) Xiaomi · MIT	1342+8/-8	6,929
30	2732	gpt-5.2-codex OpenAI · Proprietary	1341+8/-8	6,476
31	2833	kimi-k2-thinking-turbo Moonshot · Modified MIT	1331+7/-7	13,314
32	2835	gpt-5.1-codex OpenAI · Proprietary	1330+9/-9	6,508
33	3136	deepseek-v3.2 DeepSeek · MIT	1322+8/-8	7,910
34	3236	minimax-m2 MiniMax · Apache 2.0	1313+9/-9	8,833
35	3336	claude-haiku-4-5-20251001 Anthropic · Proprietary	1310+7/-7	14,613
36	3237	mimo-v2-flash (thinking) Xiaomi · MIT	1308+13/-13	2,146
37	3638	deepseek-v3.2-exp DeepSeek · MIT	1287+10/-10	5,129
38	3738	qwen3-coder-480b-a35b-instruct Alibaba · Apache 2.0	1282+7/-7	14,283
39	3944	KAT-Coder-Pro-V1 KwaiKAT · Proprietary	1260+15/-15	1,954
40	3945	qwen3.5-35b-a3b Alibaba · Apache 2.0	1257+16/-16	1,698
41	3945	gemini-3.1-flash-lite-preview Google · Proprietary	1255+17/-17	1,391
42	3945	gpt-5.1-codex-mini OpenAI · Proprietary	1244+17/-17	1,534
43	3945	qwen3.5-flash Alibaba · Proprietary	1243+17/-17	1,486
44	3945	grok-4-1-fast-reasoning xAI · Proprietary	1236+9/-9	7,131
45	4048	mistral-large-3 Mistral · Apache 2.0	1224+20/-20	1,039
46	4548	gemini-2.5-pro Google · Proprietary	1206+13/-13	3,453
47	4548	grok-4.1-thinking xAI · Proprietary	1205+19/-19	1,267
48	4548	devstral-2 Mistral · Modified MIT	1199+16/-16	1,687
49	4950	grok-4-fast-reasoning xAI · Proprietary	1154+22/-22	968
50	4951	grok-code-fast-1 xAI · Proprietary	1142+21/-21	1,017
51	5051	devstral-medium-2507 Mistral · Proprietary	1100+22/-22	1,020

Code Arena | Overall

Remove Style Control Leaderboard Plots

Fraction of Model A Wins for All Non-tied A vs. B Battles

Confidence Intervals on Model Strength (via Bootstrapping)

Battle Count for Each Combination of Models (without Ties)

Average Win Rate Against All Other Models (Uniform Sampling and No Ties)