Arena Text — Multi-Turn

Crowdsourced pairwise human preference restricted to multi-turn conversations (style control on); Bradley-Terry rating.

211models

–Open Index weight

–last updated

LMArenapublisher

#	Model	Effort	Score	Cost / task	Harness
1	Gemini 4	high	1553	–	–
2	Opus 4.6	high	1518	–	–
3	Fable 5	high	1516	–	–
4	Opus 4.7	high	1516	–	–
5	Muse Spark 1.2	xhigh	1506	–	–
6	Gemini 3.8 Flash	high	1503	–	–
7	Kimi K3	max	1500	–	–
8	Opus 5.5	high	1498	–	–
9	Gemini 3.1 Pro	high	1497	–	–
10	Muse Spark 1.1	xhigh	1497	–	–
11	Opus 4.8	high	1497	–	–
12	Gemini 3 Pro	high	1496	–	–
13	GPT-5.2	xhigh	1495	–	–
14	GPT-5.4	high	1493	–	–
15	Qwen 3.8 Max	max	1492	–	–
16	Gemini 3.7 Flash	high	1492	–	–
17	GPT-5.6 Sol	xhigh	1488	–	–
18	GPT-5.5	high	1487	–	–
19	Gemini 3.6 Flash	high	1487	–	–
20	Muse Spark 1.3	max	1487	–	–
21	GPT-6.1 Sol	max	1487	–	–
22	Opus 5	max	1486	–	–
23	Opus 4.5	high	1486	–	–
24	Fable 5.1	max	1485	–	–
25	GPT-6 Astra	max	1484	–	–
26	Sonnet 5.5	xhigh	1484	–	–
27	Grok 4.20	xhigh	1483	–	–
28	Qwen 3.7 Max	max	1483	–	–
29	GPT-5.5 Instant	–	1482	–	–
30	Gemini 3 Flash	high	1482	–	–
31	MiMo V2.5 Pro	xhigh	1481	–	–
32	Sonnet 4.6	max	1481	–	–
33	GLM 5.3	max	1481	–	–
34	Gemini 3.5 Flash	high	1480	–	–
35	GLM 5.1	max	1478	–	–
36	Sonnet 4.5	high	1478	–	–
37	Qwen 3.5 Max	–	1477	–	–
38	GLM 5.2	max	1475	–	–
39	DeepSeek V4 Pro	max	1475	–	–
40	Sonnet 5	high	1473	–	–
41	Grok 4.5	high	1472	–	–
42	GLM 5.3 Flash	max	1472	–	–
43	Opus 4.1	high	1471	–	–
44	GLM 5	high	1471	–	–
45	GPT-5.6 Terra	xhigh	1469	–	–
46	Qwen 3.6 Max	max	1469	–	–
47	GPT-6 Sol	max	1467	–	–
48	GPT-5.3	xhigh	1467	–	–
49	Gemini 3.5 Flash-Lite	xhigh	1467	–	–
50	GPT-4.5	none	1466	–	–
51	GPT-5.4 Mini	high	1466	–	–
52	DeepSeek V4.1 Flash	max	1466	–	–
53	Gemma 4 31B	max	1466	–	–
54	MiMo V2 Pro	–	1465	–	–
55	GPT-5.1	high	1464	–	–
56	Qwen 3.7 Plus	high	1464	–	–
57	Kimi K2.6	xhigh	1461	–	–
58	Grok 4.1	–	1461	–	–
59	GPT-5.6 Luna	xhigh	1457	–	–
60	GLM 4.7	none	1457	–	–
61	MiMo V2.6 Pro	high	1456	–	–
62	DeepSeek V4 Flash	max	1453	–	–
63	Qwen 3.5 397B-A17B	max	1453	–	–
64	Kimi K2.5	high	1452	–	–
65	MiMo V2	–	1452	–	–
66	MiMo V2.6 Flash	high	1452	–	–
67	MiniMax M3	max	1452	–	–
68	Hy3	xhigh	1451	–	–
69	MiMo V2.5	xhigh	1451	–	–
70	GPT-5	high	1449	–	–
71	Gemma 4 26B	xhigh	1448	–	–
72	Gemini 2.5 Pro	high	1448	–	–
73	Grok 4.6	high	1447	–	–
74	Qwen 3 Max	max	1447	–	–
75	Grok 4.3	xhigh	1446	–	–
76	Qwen 3.6 Plus	max	1445	–	–
77	GPT-6 Luna	max	1444	–	–
78	Qwen 3.8 27B	max	1441	–	–
79	Inkling	max	1441	–	–
80	Opus 4	medium	1439	–	–
81	Qwen 3 235B-A22B	xhigh	1436	–	–
82	Gemini 3.1 Flash-Lite	high	1436	–	–
83	GLM 5 V	high	1433	–	–
84	Kimi K2	medium	1433	–	–
85	Mistral Medium 3.5	high	1432	–	–
86	GPT-4.1	high	1430	–	–
87	Grok 4.7	xhigh	1430	–	–
88	DeepSeek V3.2	high	1428	–	–
89	MiniMax M2.7	high	1426	–	–
90	Mistral Large 4	high	1425	–	–
91	Haiku 4.5	high	1425	–	–
92	Qwen 3	high	1425	–	–
93	Grok 4 Fast	none	1422	–	–
94	GLM 4.6	none	1422	–	–
95	Sonnet 4	high	1421	–	–
96	o3	high	1420	–	–
97	Mistral Large 3	none	1420	–	–
98	Kimi K2.0711	–	1419	–	–
99	Qwen 3.5 122B-A10B	none	1419	–	–
100	Inkling Small	xhigh	1416	–	–
101	DeepSeek V3.1	none	1415	–	–
102	Qwen 3.5 27B	high	1414	–	–
103	Grok 4.1 Fast	high	1413	–	–
104	GPT-5.4 Nano	high	1413	–	–
105	Grok 4.0709	–	1413	–	–
106	DeepSeek V3.0324	–	1411	–	–
107	Grok 3	none	1410	–	–
108	DeepSeek R1	xhigh	1410	–	–
109	Mistral Medium 2508	–	1407	–	–
110	Kimi K2.0905	–	1404	–	–
111	DeepSeek R1.0528	–	1404	–	–
112	GLM 4.5	xhigh	1404	–	–
113	Gemini 2.5 Flash	medium	1403	–	–
114	Mistral Medium 2505	–	1403	–	–
115	Sonnet 3.7	medium	1400	–	–
116	Nemotron 3 Ultra	max	1397	–	–
117	Qwen 3 Coder	medium	1397	–	–
118	MiniMax M2.5	high	1396	–	–
119	Qwen 3.5 Flash	max	1396	–	–
120	Qwen 3.5 35B-A3B	low	1394	–	–
121	GPT-4.1 Mini	high	1393	–	–
122	MiniMax M2.1	–	1393	–	–
123	o1	high	1390	–	–
124	Sonnet 3.5	medium	1390	–	–
125	MiMo V2 Flash	high	1388	–	–
126	o4 mini	high	1383	–	–
127	Qwen 3 30B-A3B	none	1379	–	–
128	DeepSeek V3	none	1373	–	–
129	Gemini 2.5 Flash Lite	none	1373	–	–
130	Qwen 2.5 Max	none	1372	–	–
131	GPT-5 Mini	high	1372	–	–
132	GLM 4.6 V	low	1368	–	–
133	GLM 4.5 Air	–	1368	–	–
134	Mistral Small 2506	–	1365	–	–
135	GLM 4.7 Flash	none	1359	–	–
136	MiniMax M2	–	1358	–	–
137	Gemma 3 27B IT	medium	1358	–	–
138	MiniMax M1	–	1357	–	–
139	Gemini 2.0 Flash	medium	1357	–	–
140	GLM 4.5 V	none	1356	–	–
141	Grok 3 Mini	high	1350	–	–
142	Nemotron 3.5 Lightning	–	1348	–	–
143	o3 mini	high	1345	–	–
144	Gemma 3 12B IT	–	1344	–	–
145	Mercury 2	high	1343	–	–
146	Llama 3.1	max	1342	–	–
147	Llama 3.3	none	1341	–	–
148	GPT-4o	high	1341	–	–
149	Llama 3.1 405b	none	1340	–	–
150	Qwen 3 32B	none	1337	–	–
151	GLM 4	–	1333	–	–
152	Gemini 1.5 Pro	none	1329	–	–
153	Haiku 3.5	medium	1328	–	–
154	Gemini 2.0 Flash-Lite	none	1327	–	–
155	gpt-oss-120b	high	1326	–	–
156	Llama 4 Maverick	medium	1326	–	–
157	Grok 2.2024	–	1324	–	–
158	GPT-5 Nano	high	1324	–	–
159	o1 mini	high	1323	–	–
160	DeepSeek V2.5	none	1323	–	–
161	Opus 3	none	1322	–	–
162	Llama 4 Scout	medium	1319	–	–
163	Llama 3.3 70b	none	1317	–	–
164	GPT-4.1 Nano	high	1314	–	–
165	Magistral Medium 2506	–	1313	–	–
166	GPT-4	high	1313	–	–
167	GPT-4o mini	high	1311	–	–
168	Qwen 2.5 Plus	–	1307	–	–
169	GPT-4.1106	–	1299	–	–
170	Qwen 2.5 72B	none	1299	–	–
171	Mistral Large 2407	–	1297	–	–
172	Grok 2 Mini	–	1296	–	–
173	Mistral Large 2411	–	1294	–	–
174	GPT-4.0125	–	1292	–	–
175	Mistral Small 3.1	none	1291	–	–
176	gpt-oss-20b	high	1289	–	–
177	Llama 3.1 70b	none	1289	–	–
178	Gemma 2 27B IT	–	1280	–	–
179	Sonnet 3	none	1278	–	–
180	Gemini 1.5 Flash	none	1277	–	–
181	GPT-4.0314	–	1274	–	–
182	Llama 3 70b	none	1273	–	–
183	Gemma 3 4B IT	–	1272	–	–
184	Gemma 2 9B IT	–	1272	–	–
185	GPT-4.0613	–	1262	–	–
186	GLM 4.0520	–	1259	–	–
187	Qwen 2.5 Coder	none	1251	–	–
188	Mistral Small	none	1250	–	–
189	Haiku 3	medium	1245	–	–
190	Qwen 2 72B	none	1244	–	–
191	Mistral Large 2402	–	1227	–	–
192	Qwen 1.5 72B	–	1218	–	–
193	Qwen 1.5 110B	none	1213	–	–
194	Llama 3 8b	none	1206	–	–
195	GPT-3.5	high	1201	–	–
196	Llama 3.1 8b	medium	1198	–	–
197	Mistral Medium	none	1196	–	–
198	Qwen 1.5 32B	–	1193	–	–
199	Qwen 1.5 14B	–	1166	–	–
200	Gemma 2 2B IT	–	1164	–	–
201	Llama 3.2 3b	none	1149	–	–
202	Mistral	none	1136	–	–
203	Llama 2 70b	none	1135	–	–
204	Gemma 1.1 7B IT	–	1124	–	–
205	Qwen 1.5 7B	–	1114	–	–
206	Llama 2 13b	none	1101	–	–
207	Llama 2 7b	none	1078	–	–
208	Llama 3.2 1b	none	1076	–	–
209	Qwen 1.5 4B	–	1056	–	–
210	Gemma 1.1 2B IT	–	1050	–	–
211	Llama 1 3b	–	892	–	–