Frontier LLM Comparison

All major frontier large-language models, ranked by Arena Elo, context window, and output speed — across the labs that ship them

Models

Avg Elo

1369

Avg Context

800K

Avg Speed (t/s)

5.2K

Rank	Model	Lab / Company	Class	Arena Elo ▼	Context	Speed	HQ
🥇	Claude Opus 4.7 Claude Opus 4.7	Anthropic	Frontier · Reasoning · Coding	1510	200K	60	🇺🇸
🥈	Claude Opus 4.6 (Thinking) Claude Opus 4.6 (思考模式)	Anthropic	Frontier · Coding Champion	1504	200K	55	🇺🇸
🥉	GPT-5.5 GPT-5.5	OpenAI	Frontier · Agentic	1495	400K	120	🇺🇸
4	Gemini 3.1 Pro Gemini 3.1 Pro	Google DeepMind	Frontier · Multimodal · Reasoning	1493	1000K	95	🇺🇸
5	Grok 4.20 Grok 4.20	xAI	Frontier · Real-time	1491	256K	88	🇺🇸
6	GPT-5.4 (High) GPT-5.4 (高推理)	OpenAI	Frontier · Reasoning	1484	400K	88	🇺🇸
7	Claude Sonnet 4.6 Claude Sonnet 4.6	Anthropic	Workhorse · Coding	1465	1000K	115	🇺🇸
8	GLM-5.1 GLM-5.1	Z.ai (Zhipu)	Open-Weights · Coding	1465	128K	118	🇨🇳
9	ERNIE 5.0 文心一言 5.0	Baidu	Native Omni-Modal	1460	128K	108	🇨🇳
10	DeepSeek V4 Pro DeepSeek V4 Pro	DeepSeek	Open-Weights · Frontier	1455	1000K	90	🇨🇳
11	Gemini 3 Pro Gemini 3 Pro	Google DeepMind	Frontier	1450	1000K	100	🇺🇸
12	Qwen 3.6-Max-Preview 通义千问 3.6-Max-Preview	Alibaba	Closed-Weights · Coding	1448	256K	135	🇨🇳
13	Kimi K2.6 Kimi K2.6	Moonshot AI	Open-Weights · Agent · Multimodal	1442	256K	88	🇨🇳
14	GPT-5.2 GPT-5.2	OpenAI	General	1430	400K	115	🇺🇸
15	Claude Sonnet 4.5 (1M) Claude Sonnet 4.5 (1M)	Anthropic	General · 1M Context	1420	1000K	115	🇺🇸
16	o3 o3	OpenAI	Reasoning	1418	200K	42	🇺🇸
17	Llama 5 Llama 5	Meta AI	Open-Weights · Frontier	1408	5000K	62	🇺🇸
18	Mistral Large 3 Mistral Large 3	Mistral	Open-Weights · MoE	1395	128K	128	🇫🇷
19	Gemini 3.1 Flash Gemini 3.1 Flash	Google DeepMind	Fast · Multimodal	1378	1000K	240	🇺🇸
20	DeepSeek V4 Flash DeepSeek V4 Flash	DeepSeek	Open-Weights · Cheap	1370	1000K	155	🇨🇳
21	GPT-4.1 GPT-4.1	OpenAI	General · Long Context	1365	1000K	120	🇺🇸
22	DeepSeek V3.2 DeepSeek V3.2	DeepSeek	Open-Weights · MoE	1355	128K	130	🇨🇳
23	Hunyuan 3.0 混元 3.0	Tencent	In-Context Learning · Agent	1352	256K	118	🇨🇳
24	ByteDance Seed 2.0 Pro 豆包 Seed 2.0 Pro	ByteDance	Agent · Production	1340	256K	165	🇨🇳
25	Llama 4 Maverick Llama 4 Maverick	Meta AI	Open-Weights · MoE	1335	1000K	115	🇺🇸
26	Qwen3 Max 通义千问 3 Max	Alibaba	General · Multilingual	1330	262K	140	🇨🇳
27	Gemini 3.1 Flash-Lite Gemini 3.1 Flash-Lite	Google DeepMind	Cheapest · Multimodal	1318	1000K	320	🇺🇸
28	Claude Haiku 4.6 Claude Haiku 4.6	Anthropic	Cheap · Fast	1310	200K	280	🇺🇸
29	GPT-5.5 mini GPT-5.5 mini	OpenAI	Cheap · General	1305	200K	185	🇺🇸
30	o3-mini o3-mini	OpenAI	Cheap Reasoning	1295	200K	88	🇺🇸
31	Llama 4 Scout Llama 4 Scout	Meta AI	Open-Weights · 10M Context	1280	10000K	160	🇺🇸
32	DeepSeek R1 DeepSeek R1	DeepSeek	Open Reasoning	1275	128K	55	🇨🇳
33	DeepSeek-Coder V4 DeepSeek-Coder V4	DeepSeek	Coding · Open	1268	128K	128	🇨🇳
34	Mistral 3 (14B) Mistral 3 (14B)	Mistral	Open · Small Dense	1265	128K	200	🇫🇷
35	Qwen3-Coder Plus 通义千问3 Coder Plus	Alibaba	Coding · Long Context	1260	1000K	130	🇨🇳
36	Yi-Large 2 Yi-Large 2	01.AI	General	1255	200K	122	🇨🇳
37	Cohere Command A Cohere Command A	Cohere	Enterprise · RAG	1255	256K	135	🇨🇦
38	Hunyuan HY 2.0 Think 混元 HY 2.0 Think	Tencent	Reasoning · Production	1252	128K	75	🇨🇳
39	MiniMax M2.7 MiniMax M2.7	MiniMax	Open · Long Context	1245	1000K	135	🇨🇳
40	Hermes 4 Hermes 4	Nous Research	Open · Fine-tune Base	1240	128K	118	🇺🇸
41	AI21 Jamba 2 AI21 Jamba 2	AI21 Labs	Hybrid SSM · Long Context	1230	256K	148	🇮🇱