Arena WebDev · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	Opus 5.5	max	1811	–	–
2	GPT-6 Astra	max	1780	–	–
3	Sonnet 5.5	xhigh	1774	–	–
4	GPT-6.1 Sol	max	1760	–	–
5	Fable 5.1	max	1743	–	–
6	Opus 5	max	1690	–	–
7	GPT-6 Sol	max	1686	–	–
8	Gemini 4	high	1678	–	–
9	Qwen 3.8 Max	max	1673	–	–
10	Muse Spark 1.3	max	1659	–	–
11	Kimi K3	max	1655	–	–
12	Grok 4.7	xhigh	1640	–	–
13	MiMo V2.6 Flash	high	1635	–	–
14	Qwen 3.8 Flash	max	1633	–	–
15	Hy4	–	1632	–	–
16	MiMo V2.6 Pro	high	1625	–	–
17	Fable 5	high	1624	–	–
18	GLM 5.3	max	1621	–	–
19	DeepSeek V4.1 Flash	max	1619	–	–
20	Grok 4.6	high	1617	–	–
21	GPT-5.6 Sol	xhigh	1617	–	codex-harness
22	GLM 5.3 Flash	max	1606	–	–
23	GLM 5.2	max	1604	–	–
24	Qwen 3.8 27B	max	1594	–	–
25	Gemini 3.7 Flash	high	1590	–	–
26	Haiku 5.5	high	1585	–	–
27	Gemini 3.8 Flash	high	1585	–	–
28	DeepSeek V4 Pro	max	1582	–	–
29	GPT-6 Luna	max	1582	–	–
30	DeepSeek V4 Flash	high	1582	–	–
31	Opus 4.7	high	1558	–	–
32	Opus 4.8	high	1556	–	–
33	Grok 4.5	max	1553	–	–
34	Opus 4.6	high	1546	–	–
35	Mistral Large 4	high	1542	–	–
36	Muse Spark 1.1	max	1542	–	–
37	Sonnet 5	high	1540	–	–
38	Gemini 3.6 Flash	high	1538	–	–
39	Muse Spark 1.2	xhigh	1533	–	–
40	Sonnet 4.6	max	1522	–	–
41	GPT-5.6 Terra	xhigh	1522	–	codex-harness
42	GPT-5.6 Luna	xhigh	1518	–	codex-harness
43	Qwen 3.7 Max	max	1515	–	–
44	GPT-5.5	xhigh	1513	–	codex-harness
45	Kimi K2.6	max	1509	–	–
46	GLM 5.1	max	1508	–	–
47	Hy3	xhigh	1507	–	–
48	Gemini 3.5 Flash	high	1499	–	–
49	Opus 4.5	high	1493	–	–
50	MiniMax M3	max	1482	–	–
51	Qwen 3.6 Max	max	1482	–	–
52	MiMo V2.5 Pro	xhigh	1480	–	–
53	Kimi K2.7 Code	max	1473	–	–
54	GPT-5.4	high	1463	–	codex-harness
55	Qwen 3.6 Plus	max	1461	–	–
56	Gemini 3.1 Pro	high	1447	–	–
57	Gemini 3.5 Flash-Lite	xhigh	1440	–	–
58	Gemini 3 Flash	high	1440	–	–
59	Gemini 3 Pro	high	1440	–	–
60	MiMo V2.5	xhigh	1438	–	–
61	Kimi K2.5	max	1436	–	–
62	GLM 5	high	1435	–	–
63	GLM 4.7	none	1435	–	–
64	MiMo V2 Pro	–	1434	–	–
65	GPT-5	medium	1417	–	–
66	GPT-5.2	xhigh	1417	–	–
67	Inkling	max	1413	–	–
68	Inkling Small	xhigh	1408	–	–
69	GPT-5.3 Codex	xhigh	1406	–	codex-harness
70	GLM 5 V	high	1401	–	–
71	Qwen 3.5 397B-A17B	max	1400	–	–
72	MiniMax M2.7	high	1398	–	–
73	GPT-5.4 Mini	high	1398	–	–
74	GPT-5.1	medium	1395	–	–
75	Sonnet 4.5	high	1393	–	–
76	Opus 4.1	high	1390	–	–
77	MiniMax M2.5	high	1386	–	–
78	MiniMax M2.1	–	1384	–	–
79	Grok 4.20	xhigh	1375	–	–
80	Gemma 4 31B	max	1366	–	–
81	DeepSeek V3.2	high	1362	–	–
82	Qwen 3.5 122B-A10B	none	1361	–	–
83	Gemma 4 26B	xhigh	1359	–	–
84	Qwen 3.5 27B	high	1358	–	–
85	Grok 4.3	xhigh	1357	–	–
86	Laguna M.1	–	1349	–	–
87	GLM 4.6	none	1341	–	–
88	GPT-5.2 Codex	xhigh	1338	–	–
89	GPT-5.1 Codex	max	1337	–	–
90	MiMo V2 Flash	high	1330	–	–
91	Haiku 4.5	xhigh	1330	–	–
92	Kimi K2	medium	1322	–	–
93	Laguna XS.2	–	1303	–	–
94	MiniMax M2	–	1297	–	–
95	Qwen 3 Coder	medium	1275	–	–
96	Mistral Medium 3.5	high	1263	–	–
97	Gemini 3.1 Flash-Lite	high	1256	–	–
98	Qwen 3.5 35B-A3B	low	1254	–	–
99	Qwen 3.5 Flash	max	1243	–	–
100	GPT-5.1 Codex Mini	high	1243	–	–
101	Grok 4.1 Fast	high	1242	–	–
102	Mistral Large 3	none	1230	–	–
103	Gemini 2.5 Pro	high	1226	–	–
104	Grok 4.1	–	1214	–	–
105	Devstral 2	none	1196	–	–
106	Mercury 2	high	1171	–	–
107	Grok 4 Fast	none	1159	–	–
108	Devstral Medium 2507	–	1071	–	–