Chatbot Arena

This leaderboard is based on the following benchmarks.

• Chatbot Arena - a crowdsourced, randomized battle platform for large language models (LLMs). We use 2.6M+ user votes to compute Elo ratings.
• MMLU - a test to measure a model’s multitask accuracy on 57 tasks.
• Arena-Hard-Auto - an automatic evaluation tool for instruction-tuned LLMs.

| Vote | Blog | GitHub | Paper | Dataset | Twitter | Discord |

Best Open LM

Model	Arena Elo	MMLU	License
DeepSeek-R1	1361	90.8	MIT
DeepSeek-V3	1317	88.5	DeepSeek
Qwen2.5-72B-Instruct	1257	86.8	Qwen
Llama-3.3-70B-Instruct	1255	86	Llama 3.3

Full Leaderboard

Model	Arena Elo	Coding	Vision	Arena Hard	MMLU	Votes	Organization	License
🥇🏆chocolate (Early Grok-3)	1402	1399				7829	xAI	Proprietary
🥈 Gemini-2.0-Flash-Thinking-Exp-01-21	1385	1368	1280			13336	Google	Proprietary
🥈 Gemini-2.0-Pro-Exp-02-05	1379	1372	1252			11197	Google	Proprietary
🥈 ChatGPT-4o-latest (2025-01-29)	1377	1360	1276			10529	OpenAI	Proprietary
🥈 DeepSeek-R1	1361	1362				5079	DeepSeek	MIT
🥈 Gemini-2.0-Flash-001	1356	1353	1243			9092	Google	Proprietary
🥈 o1-2024-12-17	1353	1363		90.4		15437	OpenAI	Proprietary
🥉 Qwen2.5-Max	1332	1335				7370	Alibaba	Proprietary
🥉 DeepSeek-V3	1317	1317				17717	DeepSeek	DeepSeek
🥉 Qwen2.5-Plus-0125	1313	1316				3682	Alibaba	Proprietary
🥉 Gemini-2.0-Flash-Lite-Preview-02-05	1310	1322	1153			8465	Google	Proprietary
🥉 GLM-4-Plus-0111	1308	1291				4171	Zhipu	Proprietary
🥉 o3-mini	1305	1355				9338	OpenAI	Proprietary
🥉 o1-mini	1304	1353		92		55053	OpenAI	Proprietary
🥉 Step-2-16K-Exp	1304	1296				5133	StepFun	Proprietary
🥉 Gemini-1.5-Pro-002	1302	1290	1221			53271	Google	Proprietary
🥉 Grok-2-08-13	1288	1282				67117	xAI	Proprietary
🥉 Yi-Lightning	1287	1303		81.5		28955	01 AI	Proprietary
🥉 Claude 3.5 Sonnet (20241022)	1284	1325	1184	85.2	88.7	54384	Anthropic	Proprietary
Deepseek-v2.5-1210	1279	1297				7247	DeepSeek	DeepSeek
Athene-v2-Chat-72B	1275	1300		85		26182	NexusFlow	NexusFlow
GPT-4o-mini-2024-07-18	1273	1284	1124	74.94	82	64411	OpenAI	Proprietary
Gemini-1.5-Flash-002	1271	1254	1205			37039	Google	Proprietary
Llama-3.1-405B-Instruct	1269	1280			88.6	29561	Meta	Llama 3.1
Llama-3.1-Nemotron-70B-Instruct	1268	1271		84.9		7598	Nvidia	Llama 3.1
Grok-2-Mini-08-13	1266	1262				55479	xAI	Proprietary
Yi-Lightning-lite	1264	1267				17062	01 AI	Proprietary
Qwen2.5-72B-Instruct	1257	1283		78		41575	Alibaba	Qwen
GPT-4-Turbo-2024-04-09	1256	1263	1151	82.63		102116	OpenAI	Proprietary
Llama-3.3-70B-Instruct	1255	1258				23945	Meta	Llama-3.3
Mistral-Large-2407	1252	1269		70.42		48191	Mistral	Mistral Research
GPT-4-1106-preview	1250	1253				103732	OpenAI	Proprietary
Athene-70B	1250	1253		77.6		20602	NexusFlow	CC-BY-NC-4.0
Llama-3.1-70B-Instruct	1248	1251		55.73	86	58756	Meta	Llama 3.1
Claude 3 Opus	1247	1250	1076	60.36	86.8	202670	Anthropic	Proprietary
Mistral-Large-2411	1246	1265				18879	Mistral	MRL
Amazon Nova Pro 1.0	1246	1260	1044			17108	Amazon	Proprietary
GPT-4-0125-preview	1245	1244		77.96		97040	OpenAI	Proprietary
Llama-3.1-Tulu-3-70B	1244	1231				3016	Ai2	Llama 3.1
Yi-Large-preview	1240	1245		71.48		51649	01 AI	Proprietary
Claude 3.5 Haiku (20241022)	1236	1263				19050	Anthropic	Propretary
Reka-Core-20240904	1235	1222				7933	Reka AI	Proprietary
Reka-Core-20240722	1231	1208				13288	Reka AI	Proprietary
Qwen-Plus-0828	1227	1245				14612	Alibaba	Proprietary
Gemini-1.5-Flash-001	1227	1232	1072	49.61	78.9	65662	Google	Proprietary
Jamba-1.5-Large	1221	1228			81.2	9123	AI21 Labs	Jamba Open
Deepseek-v2-API-0628	1220	1242				19500	DeepSeek AI	DeepSeek
Gemma-2-27B-it	1220	1210		57.51		76608	Google	Gemma license
Qwen2.5-Coder-32B-Instruct	1217	1261				5721	Alibaba	Apache 2.0
Amazon Nova Lite 1.0	1217	1234	1060			15012	Amazon	Proprietary
Mistral-Small-24B-Instruct-2501	1217	1233				3481	Mistral	Apache 2.0
Gemma-2-9B-it-SimPO	1216	1196				10553	Princeton	MIT
Command R+ (08-2024)	1215	1181				10538	Cohere	CC-BY-NC-4.0
Deepseek-Coder-v2-0724	1214	1266		62.3		11724	DeepSeek	Proprietary
Yi-Large	1213	1220		63.7		16632	01 AI	Proprietary
Gemini-1.5-Flash-8B-001	1213	1208	1106			37728	Google	Proprietary
Llama-3.1-Nemotron-51B-Instruct	1211	1211				3895	Nvidia	Llama 3.1
Nemotron-4-340B-Instruct	1209	1198				20605	Nvidia	NVIDIA Open Model
Aya-Expanse-32B	1209	1192				28806	Cohere	CC-BY-NC-4.0
Gemini App (2024-01-24)	1208	1171				11822	Google	Proprietary
GLM-4-0520	1207	1216		63.84		10215	Zhipu AI	Proprietary
Llama-3-70B-Instruct	1207	1200		46.57	82	163746	Meta	Llama 3
Phi-4	1205	1234				11133	Microsoft	MIT
Reka-Flash-20240904	1205	1191				8125	Reka AI	Proprietary
Gemini-1.5-Flash-8B-Exp-0827	1205	1189	1112			25356	Google	Proprietary
Claude 3 Sonnet	1201	1213	1048	46.8	79	113001	Anthropic	Proprietary
Reka-Flash-20240722	1201	1187				13727	Reka AI	Proprietary
Reka-Core-20240501	1200	1190	1015		83.2	62561	Reka AI	Proprietary
Amazon Nova Micro 1.0	1198	1210				15029	Amazon	Proprietary
Gemma-2-9B-it	1192	1173				54333	Google	Gemma license
Command R+ (04-2024)	1190	1164		33.07		80864	Cohere	CC-BY-NC-4.0
Hunyuan-Standard-256K	1189	1227				2898	Tencent	Proprietary
Qwen2-72B-Instruct	1187	1187		46.86	84.2	38877	Alibaba	Qianwen LICENSE
GPT-4-0314	1186	1196		50	86.4	55947	OpenAI	Proprietary
Llama-3.1-Tulu-3-8B	1185	1178				3071	Ai2	Llama 3.1
GLM-4-0116	1183	1191		55.72		7575	Zhipu AI	Proprietary
Qwen-Max-0428	1183	1190				25678	Alibaba	Proprietary
Ministral-8B-2410	1182	1201				5108	Mistral	MRL
Aya-Expanse-8B	1180	1164				10469	Cohere	CC-BY-NC-4.0
Command R (08-2024)	1180	1162				10843	Cohere	CC-BY-NC-4.0
Claude 3 Haiku	1179	1189	1000	41.47	75.2	122288	Anthropic	Proprietary
DeepSeek-Coder-V2-Instruct	1178	1239				15757	DeepSeek AI	DeepSeek License
Llama-3.1-8B-Instruct	1176	1186		21.34	73	52635	Meta	Llama 3.1
Jamba-1.5-Mini	1176	1181			69.7	9271	AI21 Labs	Jamba Open
Reka-Flash-Preview-20240611	1165	1155	1024			20410	Reka AI	Proprietary
GPT-4-0613	1163	1167		37.9		91617	OpenAI	Proprietary
Qwen1.5-110B-Chat	1161	1175			80.4	27448	Alibaba	Qianwen LICENSE
Mistral-Large-2402	1157	1170		37.71	81.2	64904	Mistral	Proprietary
Yi-1.5-34B-Chat	1157	1162			76.8	25130	01 AI	Apache-2.0
Reka-Flash-21B-online	1156	1147				16025	Reka AI	Proprietary
QwQ-32B-Preview	1153	1146				3413	Alibaba	Apache 2.0
Llama-3-8B-Instruct	1152	1146		20.56	68.4	109223	Meta	Llama 3
InternLM2.5-20B-chat	1149	1158				10591	InternLM	Other
Claude-1	1149	1136			77	21159	Anthropic	Proprietary
Command R (04-2024)	1149	1123		17.02		56361	Cohere	CC-BY-NC-4.0
Mixtral-8x22b-Instruct-v0.1	1148	1153		36.36	77.8	53777	Mistral	Apache 2.0
Mistral Medium	1148	1152		31.9	75.3	35554	Mistral	Proprietary
Reka-Flash-21B	1148	1141			73.5	25808	Reka AI	Proprietary
Qwen1.5-72B-Chat	1147	1160		36.12	77.5	40652	Alibaba	Qianwen LICENSE
Granite-3.1-8B-Instruct	1143	1172				3303	IBM	Apache 2.0
Gemma-2-2b-it	1143	1108			51.3	46048	Google	Gemma license
Claude-2.0	1132	1135		23.99	78.5	12761	Anthropic	Proprietary
Gemini-1.0-Pro-001	1131	1103			71.8	18788	Google	Proprietary
Zephyr-ORPO-141b-A35b-v0.1	1127	1124				4861	HuggingFace	Apache 2.0
Qwen1.5-32B-Chat	1125	1149			73.4	22762	Alibaba	Qianwen LICENSE
Mistral-Next	1124	1132		27.37		12374	Mistral	Proprietary
Phi-3-Medium-4k-Instruct	1123	1125		33.37	78	26105	Microsoft	MIT
Granite-3.1-2B-Instruct	1119	1146				3386	IBM	Apache 2.0
Starling-LM-7B-beta	1119	1129		23.01		16669	Nexusflow	Apache-2.0
Claude-2.1	1118	1132		22.77		37693	Anthropic	Proprietary
GPT-3.5-Turbo-0613	1117	1135		24.82		38944	OpenAI	Proprietary
Mixtral-8x7B-Instruct-v0.1	1114	1114		23.4	70.6	76125	Mistral	Apache 2.0
Claude-Instant-1	1111	1109			73.4	20619	Anthropic	Proprietary
Yi-34B-Chat	1111	1106		23.15	73.5	15921	01 AI	Yi License
Gemini Pro	1111	1092		17.8	71.8	6559	Google	Proprietary
Qwen1.5-14B-Chat	1109	1126			67.6	18670	Alibaba	Qianwen LICENSE
GPT-3.5-Turbo-0125	1106	1124		23.34		68857	OpenAI	Proprietary
GPT-3.5-Turbo-0314	1106	1115		18.05	70	5645	OpenAI	Proprietary
WizardLM-70B-v1.0	1106	1071			63.7	8383	Microsoft	Llama 2
DBRX-Instruct-Preview	1103	1118		24.63	73.7	33730	Databricks	DBRX LICENSE
Llama-3.2-3B-Instruct	1103	1080				8404	Meta	Llama 3.2
Phi-3-Small-8k-Instruct	1102	1107		29.77	75.7	18477	Microsoft	MIT
Tulu-2-DPO-70B	1099	1093		14.99		6661	AllenAI/UW	AI2 ImpACT Low-risk
Granite-3.0-8B-Instruct	1093	1097				7005	IBM	Apache 2.0
Llama-2-70B-chat	1093	1072		11.55	63	39616	Meta	Llama 2
OpenChat-3.5-0106	1092	1102			65.8	12977	OpenChat	Apache-2.0
Vicuna-33B	1091	1067		8.63	59.2	22950	LMSYS	Non-commercial
Snowflake Arctic Instruct	1090	1077		17.61	67.3	34168	Snowflake	Apache 2.0
Starling-LM-7B-alpha	1088	1080		12.8	63.9	10417	UC Berkeley	CC-BY-NC-4.0
Gemma-1.1-7B-it	1084	1084			64.3	25067	Google	Gemma license
Nous-Hermes-2-Mixtral-8x7B-DPO	1084	1079				3833	NousResearch	Apache-2.0
NV-Llama2-70B-SteerLM-Chat	1081	1023			68.5	3638	Nvidia	Llama 2
pplx-70B-online	1078	1028				6894	Perplexity AI	Proprietary
DeepSeek-LLM-67B-Chat	1077	1079			71.3	4986	DeepSeek AI	DeepSeek License
OpenChat-3.5	1076	1054			64.3	8106	OpenChat	Apache-2.0
Granite-3.0-2B-Instruct	1074	1088				7186	IBM	Apache 2.0
OpenHermes-2.5-Mistral-7B	1074	1058				5086	NousResearch	Apache-2.0
Mistral-7B-Instruct-v0.2	1072	1074				20048	Mistral	Apache-2.0
Phi-3-Mini-4K-Instruct-June-24	1071	1082			70.9	12811	Microsoft	MIT
Qwen1.5-7B-Chat	1070	1089			61	4866	Alibaba	Qianwen LICENSE
GPT-3.5-Turbo-1106	1068	1095				17028	OpenAI	Proprietary
Phi-3-Mini-4k-Instruct	1066	1086			68.8	21089	Microsoft	MIT
Llama-2-13b-chat	1063	1051			53.6	19732	Meta	Llama 2
SOLAR-10.7B-Instruct-v1.0	1062	1047			66.2	4290	Upstage AI	CC-BY-NC-4.0
Dolphin-2.2.1-Mistral-7B	1062	1024				1713	Cognitive Computations	Apache-2.0
WizardLM-13b-v1.2	1059	1026			52.7	7175	Microsoft	Llama 2
Llama-3.2-1B-Instruct	1054	1047				8528	Meta	Llama 3.2
Step-1o-Vision-32k (highres)			1180				StepFun	Proprietary
Qwen2.5-VL-72B-Instruct			1164				Alibaba	Qwen
Pixtral-Large-2411			1153				Mistral	MRL
Qwen-VL-Max-1119			1127				Alibaba	Proprietary
Step-1V-32K			1111				StepFun	Proprietary
Qwen2-VL-72b-Instruct			1110				Alibaba	Qwen
Molmo-72B-0924			1075				AI2	Apache 2.0
Pixtral-12B-2409			1072				Mistral	Apache 2.0
Llama-3.2-90B-Vision-Instruct			1069				Meta	Llama 3.2
InternVL2-26B			1067				OpenGVLab	MIT
Qwen2-VL-7B-Instruct			1053				Aliaba	Apache 2.0
Yi-Vision			1045				01 AI	Proprietary
Llama-3.2-11B-Vision-Instruct			1032				Meta	Llama 3.2

If you want to see more models, please help us add them.

💻 Code: The Arena Elo ratings are computed by this notebook. The MT-bench scores (single-answer grading on a scale of 10) are computed by fastchat.llm_judge. The MMLU scores are computed by InstructEval. Higher values are better for all benchmarks. Empty cells mean not available. The latest and detailed leaderboard is here.

More Statistics for Chatbot Arena

🔗 Arena Statistics

Transition from online Elo rating system to Bradley-Terry model

We adopted the Elo rating system for ranking models since the launch of the Arena. It has been useful to transform pairwise human preference to Elo ratings that serve as a predictor of winrate between models. Specifically, if player A has a rating of R_A and player B a rating of R_B, the probability of player A winning is

ELO rating has been used to rank chess players by the international community for over 60 years. Standard Elo rating systems assume a player’s performance changes overtime. So an online algorithm is needed to capture such dynamics, meaning recent games should weigh more than older games. Specifically, after each game, a player’s rating is updated according to the difference between predicted outcome and actual outcome.

This algorithm has two distinct features:

1. It can be computed asynchronously by players around the world.
2. It allows for players performance to change dynamically – it does not assume a fixed unknown value for the players rating.

This ability to adapt is determined by the parameter K which controls the magnitude of rating changes that can affect the overall result. A larger K essentially put more weight on the recent games, which may make sense for new players whose performance improves quickly. However as players become more senior and their performance “converges” then a smaller value of K is more appropriate. As a result, USCF adopted K based on the number of games and tournaments completed by the player (reference). That is, the Elo rating of a senior player changes slower than a new player.

When we launched the Arena, we noticed considerable variability in the ratings using the classic online algorithm. We tried to tune the K to be sufficiently stable while also allowing new models to move up quickly in the leaderboard. We ultimately decided to adopt a bootstrap-like technique to shuffle the data and sample Elo scores from 1000 permutations of the online plays. You can find the details in this notebook. This provided consistent stable scores and allowed us to incorporate new models quickly. This is also observed in a recent work by Cohere. However, we used the same samples to estimate confidence intervals which were therefore too wide (effectively CI’s for the original online Elo estimates).

In the context of LLM ranking, there are two important differences from the classic Elo chess ranking system. First, we have access to the entire history of all games for all models and so we don’t need a decentralized algorithm. Second, most models are static (we have access to the weights) and so we don’t expect their performance to change. However, it is worth noting that the hosted proprietary models may not be static and their behavior can change without notice. We try our best to pin specific model API versions if possible.

To improve the quality of our rankings and their confidence estimates, we are adopting another widely used rating system called the Bradley–Terry (BT) model. This model actually is the maximum likelihood (MLE) estimate of the underlying Elo model assuming a fixed but unknown pairwise win-rate. Similar to Elo rating, BT model is also based on pairwise comparison to derive ratings of players to estimate win rate between each other. The core difference between BT model vs the online Elo system is the assumption that player’s performance does not change (i.e., game order does not matter) and the computation takes place in a centralized fashion.