LMArena (the project formerly known as LMSys Chatbot Arena) is a public benchmark where humans blind-vote between two anonymous model responses to the same prompt. Votes are aggregated into Elo ratings — the same statistical method used to rank chess players — so a model rating only goes up if it consistently beats peers in side-by-side comparison.

Who is currently #1 on LMArena in May 2026?

Our May 2026 snapshot of the LMArena text leaderboard puts OpenAI: GPT-5.4 Pro at the top with undefined Elo. The full top 10 is below; the rankings re-shuffle weekly as votes accumulate.

How should I read Arena Elo as an enterprise buyer?

Arena Elo is a strong signal for general chat quality and human preference under short-conversation conditions. It does not predict your production accuracy on function-calling, long-context retrieval, or domain-specific tasks. Use it as a triage filter, not a procurement decision — then build an internal eval harness on your own workload.

Why are there separate LMArena leaderboards?

The Arena now publishes separate leaderboards for text chat, coding, hard prompts, math, multi-turn, vision, and several language-specific variants. A model can lead one and rank mid-pack on another. The split between text and coding leaders is the most cited illustration.

Is the LMArena gameable?

Partially. Style polish and short-answer formatting tend to win votes, which can favour models tuned for chat over models tuned for accuracy. The Arena team has shipped category-specific leaderboards and prompt-category filters to reduce this skew. It is still the best human-preference signal we have at scale.

Updated Apr 6, 2026

LMArena Leaderboard — May 2026

What the LMArena actually is, how to read an Arena Elo score, and the current top 10 for May 2026. The original human-preference benchmark that started as LMSys Chatbot Arena and now anchors most enterprise model selection conversations.

What is the LMArena, in one paragraph?

The LMArena is a public, blind side-by-side voting site for AI chat models. A user submits a prompt, two anonymous models reply, the user picks a winner, and the project aggregates millions of such votes into Elo ratings. It started in 2023 as the LMSys Chatbot Arena out of UC Berkeley and rebranded to LMArena.ai in 2024-25 as it spun out into an independent project. The current May 2026 top 10 is below — three models now sit above the historical 1500 Elo barrier on text, with the open-weights tier within striking distance of the closed-source frontier.

How to read an Arena Elo score

Reference table for Arena Elo bands (May 2026)

  1500+   Frontier         Gemini 3.1 Pro, Claude 4.7, GPT-5.5 Pro
  1450    Frontier-adj.    DeepSeek V4 Pro, Qwen 3.6 Plus
  1400    Strong tier      GPT-4.1, Claude Sonnet 4, Gemini 2.5 Pro
  1300    Capable tier     Llama 4 Maverick, Mistral Large 3
  1200    Solid daily      Gemma 4, Phi-4, Mistral Small 3
  1100    Light tasks      DeepSeek V4 Flash, GPT-4o Mini
   <1100  Legacy tier      Older 2023-24 model generations

A 100-Elo gap means the higher-rated model wins ~64% of head-to-heads.
A 200-Elo gap means it wins ~76%. Rating shifts under 25 points are noise.

Live Leaderboard

318 models

#	Model	Quality	Arena ELO	Speed	Price	Context	Value	Released
1	OpenAI: GPT-5.4 Pro OpenAI · Complex analysis	97	—	—	$30 / $180	1M	0.9	Mar 2026
2	OpenAI: GPT-5.2 Pro OpenAI · Complex analysis	97	—	—	$21 / $168	400K	1.0	Dec 2025
3	OpenAI: o3 Deep Research OpenAI · Deep research	96	—	—	$10 / $40	200K	3.8	Oct 2025
4	OpenAI: o4 Mini Deep Research OpenAI · Deep research	96	—	—	$2 / $8	200K	19.2	Oct 2025
5	OpenAI: o3 Pro OpenAI · Hard reasoning	96	—	—	$20 / $80	200K	1.9	Jun 2025
6	Anthropic: Claude Opus 4.6 Anthropic · General purpose	95	—	—	$5 / $25	1M	6.3	Feb 2026
7	Anthropic: Claude Opus 4.5 Anthropic · General purpose	95	—	—	$5 / $25	200K	6.3	Nov 2025
8	Google: Gemini 3.1 Pro Preview Custom Tools Google · Speed & cost	94	—	—	$2 / $12	1M	13.4	Feb 2026
9	Google: Gemini 3.1 Pro Preview Google · Speed & cost	94	—	—	$2 / $12	1M	13.4	Feb 2026
10	Google: Nano Banana Pro (Gemini 3 Pro Image Preview) Google · Image generation	94	—	—	$2 / $12	66K	13.4	Nov 2025
11	Anthropic: Claude Opus 4.1 Anthropic · Multimodal	94	—	—	$15 / $75	200K	2.1	Aug 2025
12	Anthropic: Claude Opus 4 Anthropic · Multimodal	94	—	—	$15 / $75	200K	2.1	May 2025
13	OpenAI: GPT-5.4 OpenAI · General purpose	93	—	—	$2.5 / $15	1M	10.6	Mar 2026
14	OpenAI: GPT-5.3 Chat OpenAI · General purpose	93	—	—	$1.75 / $14	128K	11.8	Mar 2026
15	OpenAI: GPT-5.3-Codex OpenAI · Code generation	93	—	—	$1.75 / $14	400K	11.8	Feb 2026
16	OpenAI: GPT-5.2-Codex OpenAI · Code generation	93	—	—	$1.75 / $14	400K	11.8	Jan 2026
17	OpenAI: GPT-5.2 Chat OpenAI · General purpose	93	—	—	$1.75 / $14	128K	11.8	Dec 2025
18	OpenAI: GPT-5.2 OpenAI · General purpose	93	—	—	$1.75 / $14	400K	11.8	Dec 2025
19	OpenAI: GPT-5.1-Codex-Max OpenAI · Code generation	93	—	—	$1.25 / $10	400K	16.5	Dec 2025
20	OpenAI: GPT-5.1 OpenAI · General purpose	93	—	—	$1.25 / $10	400K	16.5	Nov 2025
21	OpenAI: GPT-5.1 Chat OpenAI · General purpose	93	—	—	$1.25 / $10	128K	16.5	Nov 2025
22	OpenAI: GPT-5.1-Codex OpenAI · Code generation	93	—	—	$1.25 / $10	400K	16.5	Nov 2025
23	OpenAI: o1-pro OpenAI · Hard reasoning	93	—	—	$150 / $600	200K	0.2	Mar 2025
24	OpenAI: GPT-4 (older v0314) OpenAI · Complex analysis	93	—	—	$30 / $60	8K	2.1	May 2023
25	OpenAI: GPT-4 OpenAI · Multimodal	93	—	—	$30 / $60	8K	2.1	May 2023
26	OpenAI: GPT-5 OpenAI · General purpose	92	—	—	$1.25 / $10	400K	16.4	Aug 2025
27	OpenAI: o3 OpenAI · Hard reasoning	92	—	—	$2 / $8	200K	18.4	Apr 2025
28	TNG: DeepSeek R1T2 ChimeraOSS · Hard reasoning	91	—	—	$0.3 / $1.1	164K	130.0	Jul 2025
29	Google: Gemini 2.5 Pro Google · Speed & cost	91	—	—	$1.25 / $10	1M	16.2	Jun 2025
30	Google: Gemini 2.5 Pro Preview 06-05 Google · Speed & cost	91	—	—	$1.25 / $10	1M	16.2	Jun 2025
31	DeepSeek: R1 0528OSS DeepSeek · Hard reasoning	91	—	—	$0.45 / $2.15	164K	70.0	May 2025
32	Google: Gemini 2.5 Pro Preview 05-06 Google · Speed & cost	91	—	—	$1.25 / $10	1M	16.2	May 2025
33	DeepSeek: R1 Distill Qwen 32BOSS DeepSeek · Hard reasoning	91	—	—	$0.29 / $0.29	33K	313.8	Jan 2025
34	DeepSeek: R1 Distill Llama 70BOSS DeepSeek · Hard reasoning	91	—	—	$0.7 / $0.8	131K	121.3	Jan 2025
35	DeepSeek: R1OSS DeepSeek · Hard reasoning	91	—	—	$0.7 / $2.5	64K	56.9	Jan 2025
36	xAI: Grok 3 xAI · General purpose	90	—	—	$3 / $15	131K	10.0	Jun 2025
37	xAI: Grok 3 Beta xAI · General purpose	90	—	—	$3 / $15	131K	10.0	Apr 2025
38	OpenAI: GPT-4.1 OpenAI · General purpose	89	—	—	$2 / $8	1M	17.8	Apr 2025
39	Anthropic: Claude Sonnet 4.6 Anthropic · General purpose	88	—	—	$3 / $15	1M	9.8	Feb 2026
40	OpenAI: GPT-5 Image OpenAI · Multimodal	88	—	—	$10 / $10	400K	8.8	Oct 2025
41	OpenAI: GPT-5 Pro OpenAI · Complex analysis	88	—	—	$15 / $120	400K	1.3	Oct 2025
42	Anthropic: Claude Sonnet 4.5 Anthropic · General purpose	88	—	—	$3 / $15	1M	9.8	Sep 2025
43	OpenAI: GPT-4o Audio OpenAI · General purpose	88	—	—	$2.5 / $10	128K	14.1	Aug 2025
44	OpenAI: GPT-4o Search Preview OpenAI · Search + citations	88	—	—	$2.5 / $10	128K	14.1	Mar 2025
45	OpenAI: o1 OpenAI · Hard reasoning	88	—	—	$15 / $60	200K	2.3	Dec 2024
46	OpenAI: GPT-4o (2024-11-20) OpenAI · General purpose	88	—	—	$2.5 / $10	128K	14.1	Nov 2024
47	OpenAI: GPT-4o (2024-08-06) OpenAI · General purpose	88	—	—	$2.5 / $10	128K	14.1	Aug 2024
48	OpenAI: GPT-4o OpenAI · General purpose	88	—	—	$2.5 / $10	128K	14.1	May 2024
49	OpenAI: GPT-4o (extended) OpenAI · Multimodal	88	—	—	$6 / $18	128K	7.3	May 2024
50	OpenAI: GPT-4o (2024-05-13) OpenAI · General purpose	88	—	—	$5 / $15	128K	8.8	May 2024
51	OpenAI: GPT-4 Turbo OpenAI · Multimodal	88	—	—	$10 / $30	128K	4.4	Apr 2024
52	OpenAI: GPT-4 Turbo Preview OpenAI · Complex analysis	88	—	—	$10 / $30	128K	4.4	Jan 2024
53	OpenAI: GPT-4 Turbo (older v1106) OpenAI · Multimodal	88	—	—	$10 / $30	128K	4.4	Nov 2023
54	Nex AGI: DeepSeek V3.1 Nex N1OSS · Open-source	86	—	—	$0.135 / $0.5	131K	270.9	Dec 2025
55	DeepSeek: DeepSeek V3.2 SpecialeOSS DeepSeek · Open-source	86	—	—	$0.4 / $1.2	164K	107.5	Dec 2025
56	DeepSeek: DeepSeek V3.2OSS DeepSeek · Open-source	86	—	—	$0.26 / $0.38	164K	268.8	Dec 2025
57	DeepSeek: DeepSeek V3.2 ExpOSS DeepSeek · Open-source	86	—	—	$0.27 / $0.41	164K	252.9	Sep 2025
58	DeepSeek: DeepSeek V3.1 TerminusOSS DeepSeek · Open-source	86	—	—	$0.21 / $0.79	164K	172.0	Sep 2025
59	DeepSeek: DeepSeek V3.1OSS DeepSeek · Open-source	86	—	—	$0.15 / $0.75	33K	191.1	Aug 2025
60	Anthropic: Claude Sonnet 4 Anthropic · General purpose	86	—	—	$3 / $15	200K	9.6	May 2025
61	DeepSeek: DeepSeek V3 0324OSS DeepSeek · Open-source	86	—	—	$0.2 / $0.77	164K	177.3	Mar 2025
62	Anthropic: Claude 3.7 Sonnet Anthropic · General purpose	86	—	—	$3 / $15	200K	9.6	Feb 2025
63	Anthropic: Claude 3.7 Sonnet (thinking) Anthropic · Hard reasoning	86	—	—	$3 / $15	200K	9.6	Feb 2025
64	DeepSeek: DeepSeek V3OSS DeepSeek · Open-source	86	—	—	$0.32 / $0.89	164K	142.1	Dec 2024
65	Mistral: Mistral Large 3 2512OSS Mistral AI · Open-source	85	—	—	$0.5 / $1.5	262K	85.0	Dec 2025
66	Mistral Large 2411OSS Mistral AI · Open-source	85	—	—	$2 / $6	131K	21.3	Nov 2024
67	Mistral Large 2407OSS Mistral AI · Open-source	85	—	—	$2 / $6	131K	21.3	Nov 2024
68	Mistral LargeOSS Mistral AI · Open-source	85	—	—	$2 / $6	128K	21.3	Feb 2024
69	Cohere: Command R+ (08-2024)OSS Cohere · Open-source	84	—	—	$2.5 / $10	128K	13.4	Aug 2024
70	OpenAI: GPT-5.4 Mini OpenAI · Speed & cost	83	—	—	$0.75 / $4.5	400K	31.6	Mar 2026
71	OpenAI: GPT-5 Mini OpenAI · Speed & cost	83	—	—	$0.25 / $2	400K	73.8	Aug 2025
72	Qwen: Qwen3.5-9BOSS Alibaba Cloud · Open-source	82	—	—	$0.05 / $0.15	256K	820.0	Mar 2026
73	Qwen: Qwen3.5-35B-A3BOSS Alibaba Cloud · Open-source	82	—	—	$0.1625 / $1.3	262K	112.1	Feb 2026
74	Qwen: Qwen3.5-27BOSS Alibaba Cloud · Open-source	82	—	—	$0.195 / $1.56	262K	93.4	Feb 2026
75	Qwen: Qwen3.5-122B-A10BOSS Alibaba Cloud · Open-source	82	—	—	$0.26 / $2.08	262K	70.1	Feb 2026
76	Qwen: Qwen3.5-FlashOSS Alibaba Cloud · Speed & cost	82	—	—	$0.065 / $0.26	1M	504.6	Feb 2026
77	Qwen: Qwen3.5 Plus 2026-02-15OSS Alibaba Cloud · Open-source	82	—	—	$0.26 / $1.56	1M	90.1	Feb 2026
78	Qwen: Qwen3.5 397B A17BOSS Alibaba Cloud · Open-source	82	—	—	$0.39 / $2.34	262K	60.1	Feb 2026
79	Qwen: Qwen3 Max ThinkingOSS Alibaba Cloud · Hard reasoning	82	—	—	$0.78 / $3.9	262K	35.0	Feb 2026
80	Qwen: Qwen3 Coder NextOSS Alibaba Cloud · Code generation	82	—	—	$0.12 / $0.75	262K	188.5	Feb 2026
81	Qwen: Qwen3 VL 32B InstructOSS Alibaba Cloud · Open-source	82	—	—	$0.104 / $0.416	131K	315.4	Oct 2025
82	Qwen: Qwen3 VL 8B ThinkingOSS Alibaba Cloud · Hard reasoning	82	—	—	$0.117 / $1.365	131K	110.7	Oct 2025
83	Qwen: Qwen3 VL 8B InstructOSS Alibaba Cloud · Open-source	82	—	—	$0.08 / $0.5	131K	282.8	Oct 2025
84	Qwen: Qwen3 VL 30B A3B ThinkingOSS Alibaba Cloud · Hard reasoning	82	—	—	$0.13 / $1.56	131K	97.0	Oct 2025
85	Qwen: Qwen3 VL 30B A3B InstructOSS Alibaba Cloud · Open-source	82	—	—	$0.13 / $0.52	131K	252.3	Oct 2025
86	Qwen: Qwen3 VL 235B A22B ThinkingOSS Alibaba Cloud · Hard reasoning	82	—	—	$0.26 / $2.6	131K	57.3	Sep 2025
87	Qwen: Qwen3 VL 235B A22B InstructOSS Alibaba Cloud · Open-source	82	—	—	$0.2 / $0.88	262K	151.9	Sep 2025
88	Qwen: Qwen3 MaxOSS Alibaba Cloud · Open-source	82	—	—	$0.78 / $3.9	262K	35.0	Sep 2025
89	Qwen: Qwen3 Coder PlusOSS Alibaba Cloud · Code generation	82	—	—	$0.65 / $3.25	1M	42.1	Sep 2025
90	Qwen: Qwen3 Coder FlashOSS Alibaba Cloud · Code generation	82	—	—	$0.195 / $0.975	1M	140.2	Sep 2025
91	Qwen: Qwen3 Next 80B A3B ThinkingOSS Alibaba Cloud · Hard reasoning	82	—	—	$0.0975 / $0.78	131K	186.9	Sep 2025
92	Qwen: Qwen3 Next 80B A3B InstructOSS Alibaba Cloud · Open-source	82	—	—	$0.09 / $1.1	262K	137.8	Sep 2025
93	Qwen: Qwen3 30B A3B Thinking 2507OSS Alibaba Cloud · Hard reasoning	82	—	—	$0.08 / $0.4	131K	341.7	Aug 2025
94	Qwen: Qwen3 Coder 30B A3B InstructOSS Alibaba Cloud · Code generation	82	—	—	$0.07 / $0.27	160K	482.4	Jul 2025
95	Qwen: Qwen3 30B A3B Instruct 2507OSS Alibaba Cloud · Open-source	82	—	—	$0.09 / $0.3	262K	420.5	Jul 2025
96	Qwen: Qwen3 235B A22B Thinking 2507OSS Alibaba Cloud · Hard reasoning	82	—	—	$0.1495 / $1.495	131K	99.7	Jul 2025
97	Qwen: Qwen3 Coder 480B A35BOSS Alibaba Cloud · Code generation	82	—	—	$0.22 / $1	262K	134.4	Jul 2025
98	Qwen: Qwen3 235B A22B Instruct 2507OSS Alibaba Cloud · Open-source	82	—	—	$0.071 / $0.1	262K	959.1	Jul 2025
99	xAI: Grok 3 Mini xAI · Speed & cost	82	—	—	$0.3 / $0.5	131K	205.0	Jun 2025
100	Qwen: Qwen3 30B A3BOSS Alibaba Cloud · Open-source	82	—	—	$0.08 / $0.28	41K	455.6	Apr 2025
101	Qwen: Qwen3 8BOSS Alibaba Cloud · Open-source	82	—	—	$0.05 / $0.4	41K	364.4	Apr 2025
102	Qwen: Qwen3 14BOSS Alibaba Cloud · Open-source	82	—	—	$0.06 / $0.24	41K	546.7	Apr 2025
103	Qwen: Qwen3 32BOSS Alibaba Cloud · Open-source	82	—	—	$0.08 / $0.24	41K	512.5	Apr 2025
104	Qwen: Qwen3 235B A22BOSS Alibaba Cloud · Open-source	82	—	—	$0.455 / $1.82	131K	72.1	Apr 2025
105	OpenAI: o4 Mini High OpenAI · Hard reasoning	82	—	—	$1.1 / $4.4	200K	29.8	Apr 2025
106	OpenAI: o4 Mini OpenAI · Hard reasoning	82	—	—	$1.1 / $4.4	200K	29.8	Apr 2025
107	xAI: Grok 3 Mini Beta xAI · Speed & cost	82	—	—	$0.3 / $0.5	131K	205.0	Apr 2025
108	Meta: Llama 4 MaverickOSS Meta · Open-source	82	—	—	$0.15 / $0.6	1M	218.7	Apr 2025
109	OpenAI: o3 Mini High OpenAI · Hard reasoning	82	—	—	$1.1 / $4.4	200K	29.8	Feb 2025
110	AionLabs: Aion-1.0 · General purpose	82	—	—	$4 / $8	131K	13.7	Feb 2025
111	OpenAI: o3 Mini OpenAI · Hard reasoning	82	—	—	$1.1 / $4.4	200K	29.8	Jan 2025
112	Qwen2.5 72B InstructOSS Alibaba Cloud · Open-source	82	—	—	$0.12 / $0.39	33K	321.6	Sep 2024
113	Goliath 120B · General purpose	82	—	—	$3.75 / $7.5	6K	14.6	Nov 2023
114	Google: Gemini 3 Flash Preview Google · Speed & cost	80	—	—	$0.5 / $3	1M	45.7	Dec 2025
115	Google: Nano Banana (Gemini 2.5 Flash Image) Google · Image generation	80	—	—	$0.3 / $2.5	33K	57.1	Oct 2025
116	Google: Gemini 2.5 Flash Lite Preview 09-2025 Google · Speed & cost	80	—	—	$0.1 / $0.4	1M	320.0	Sep 2025
117	Google: Gemini 2.5 Flash Lite Google · Speed & cost	80	—	—	$0.1 / $0.4	1M	320.0	Jul 2025
118	Google: Gemini 2.5 Flash Google · Speed & cost	80	—	—	$0.3 / $2.5	1M	57.1	Jun 2025
119	OpenAI: GPT-4.1 Mini OpenAI · Speed & cost	80	—	—	$0.4 / $1.6	1M	80.0	Apr 2025
120	OpenAI: GPT-4o-mini Search Preview OpenAI · Search + citations	80	—	—	$0.15 / $0.6	128K	213.3	Mar 2025
121	OpenAI: GPT-4o-mini (2024-07-18) OpenAI · Speed & cost	80	—	—	$0.15 / $0.6	128K	213.3	Jul 2024
122	OpenAI: GPT-4o-mini OpenAI · Speed & cost	80	—	—	$0.15 / $0.6	128K	213.3	Jul 2024
123	Mistral: Codestral 2508OSS Mistral AI · Code generation	78	—	—	$0.3 / $0.9	256K	130.0	Aug 2025
124	Google: Gemma 4 26B A4B OSS Google · Open-source	76	—	—	$0.13 / $0.4	262K	286.8	Apr 2026
125	Google: Gemma 4 31BOSS Google · Open-source	76	—	—	$0.14 / $0.4	262K	281.5	Apr 2026
126	Anthropic: Claude Haiku 4.5 Anthropic · Speed & cost	76	—	—	$1 / $5	200K	25.3	Oct 2025
127	Anthropic: Claude 3.5 Haiku Anthropic · Speed & cost	76	—	—	$0.8 / $4	200K	31.7	Nov 2024
128	Z.ai: GLM 5V TurboOSS · Open-source	74	—	—	$1.2 / $4	203K	28.5	Apr 2026
129	xAI: Grok 4.20 Multi-Agent xAI · General purpose	74	—	—	$2 / $6	2M	18.5	Mar 2026
130	xAI: Grok 4.20 xAI · General purpose	74	—	—	$2 / $6	2M	18.5	Mar 2026
131	Z.ai: GLM 5 TurboOSS · Open-source	74	—	—	$1.2 / $4	203K	28.5	Mar 2026
132	OpenAI: GPT Audio OpenAI · General purpose	74	—	—	$2.5 / $10	128K	11.8	Jan 2026
133	Deep Cogito: Cogito v2.1 671B · General purpose	74	—	—	$1.25 / $1.25	128K	59.2	Nov 2025
134	Amazon: Nova Premier 1.0 Amazon · General purpose	74	—	—	$2.5 / $12.5	1M	9.9	Oct 2025
135	Perplexity: Sonar Pro Search Perplexity · Search + citations	74	—	—	$3 / $15	200K	8.2	Oct 2025
136	OpenAI: GPT-5 Image Mini OpenAI · Image generation	74	—	—	$2.5 / $2	400K	32.9	Oct 2025
137	NVIDIA: Llama 3.3 Nemotron Super 49B V1.5OSS · Open-source	74	—	—	$0.1 / $0.4	131K	296.0	Oct 2025
138	OpenAI: GPT-5 Codex OpenAI · Code generation	74	—	—	$1.25 / $10	400K	13.2	Sep 2025
139	AI21: Jamba Large 1.7 · General purpose	74	—	—	$2 / $8	256K	14.8	Aug 2025
140	OpenAI: GPT-5 Chat OpenAI · General purpose	74	—	—	$1.25 / $10	128K	13.2	Aug 2025
141	xAI: Grok 4 xAI · General purpose	74	—	—	$3 / $15	256K	8.2	Jul 2025
142	Meta: Llama 4 ScoutOSS Meta · Speed & cost	74	—	—	$0.08 / $0.3	328K	389.5	Apr 2025
143	Google: Gemma 3 12BOSS Google · Open-source	74	—	—	$0.04 / $0.13	131K	870.6	Mar 2025
144	Cohere: Command A Cohere · General purpose	74	—	—	$2.5 / $10	256K	11.8	Mar 2025
145	Google: Gemma 3 27BOSS Google · Open-source	74	—	—	$0.08 / $0.16	131K	616.7	Mar 2025
146	Perplexity: Sonar Reasoning Pro Perplexity · Search + citations	74	—	—	$2 / $8	128K	14.8	Mar 2025
147	Perplexity: Sonar Pro Perplexity · Search + citations	74	—	—	$3 / $15	200K	8.2	Mar 2025
148	Perplexity: Sonar Deep Research Perplexity · Deep research	74	—	—	$2 / $8	128K	14.8	Mar 2025
149	Qwen: Qwen-Max OSS Alibaba Cloud · Open-source	74	—	—	$1.04 / $4.16	33K	28.5	Feb 2025
150	Sao10K: Llama 3.1 70B Hanami x1OSS · Hard reasoning	74	—	—	$3 / $3	16K	24.7	Jan 2025
151	Meta: Llama 3.3 70B InstructOSS Meta · Open-source	74	—	—	$0.1 / $0.32	131K	352.4	Dec 2024
152	Mistral: Pixtral Large 2411OSS Mistral AI · Open-source	74	—	—	$2 / $6	131K	18.5	Nov 2024
153	Magnum v4 72B · General purpose	74	—	—	$3 / $5	16K	18.5	Oct 2024
154	NVIDIA: Llama 3.1 Nemotron 70B InstructOSS · Open-source	74	—	—	$1.2 / $1.2	131K	61.7	Oct 2024
155	Inflection: Inflection 3 Pi · General purpose	74	—	—	$2.5 / $10	8K	11.8	Oct 2024
156	Inflection: Inflection 3 Productivity · General purpose	74	—	—	$2.5 / $10	8K	11.8	Oct 2024
157	Nous: Hermes 3 70B InstructOSS · Search + citations	74	—	—	$0.3 / $0.3	131K	246.7	Aug 2024
158	Meta: Llama 3.1 70B InstructOSS Meta · Open-source	74	—	—	$0.4 / $0.4	131K	185.0	Jul 2024
159	Sao10k: Llama 3 Euryale 70B v2.1OSS · Hard reasoning	74	—	—	$1.48 / $1.48	8K	50.0	Jun 2024
160	OpenAI: GPT-3.5 Turbo Instruct OpenAI · General purpose	74	—	—	$1.5 / $2	4K	42.3	Sep 2023
161	OpenAI: GPT-3.5 Turbo 16k OpenAI · General purpose	74	—	—	$3 / $4	16K	21.1	Aug 2023
162	Google: Gemini 2.0 Flash Lite Google · Speed & cost	73	—	—	$0.075 / $0.3	1M	389.3	Feb 2025
163	Google: Gemini 2.0 Flash Google · Speed & cost	73	—	—	$0.1 / $0.4	1M	292.0	Feb 2025
164	OpenAI: GPT-5.4 Nano OpenAI · Speed & cost	72	—	—	$0.2 / $1.25	400K	99.3	Mar 2026
165	Mistral: Mistral Small 4OSS Mistral AI · Open-source	72	—	—	$0.15 / $0.6	262K	192.0	Mar 2026
166	Mistral: Mistral Small CreativeOSS Mistral AI · Open-source	72	—	—	$0.1 / $0.3	33K	360.0	Dec 2025
167	OpenAI: GPT-5 Nano OpenAI · Speed & cost	72	—	—	$0.05 / $0.4	400K	320.0	Aug 2025
168	Mistral: Mistral Small 3.2 24BOSS Mistral AI · Open-source	72	—	—	$0.075 / $0.2	128K	523.6	Jun 2025
169	OpenAI: GPT-4.1 Nano OpenAI · Speed & cost	72	—	—	$0.1 / $0.4	1M	288.0	Apr 2025
170	Mistral: Mistral Small 3.1 24BOSS Mistral AI · Open-source	72	—	—	$0.03 / $0.11	131K	1028.6	Mar 2025
171	Mistral: Mistral Small 3OSS Mistral AI · Open-source	72	—	—	$0.05 / $0.08	33K	1107.7	Jan 2025
172	Mistral: Mistral NemoOSS Mistral AI · Open-source	72	—	—	$0.02 / $0.04	131K	2400.0	Jul 2024
173	Mistral: Mixtral 8x22B InstructOSS Mistral AI · Open-source	72	—	—	$2 / $6	66K	18.0	Apr 2024
174	Anthropic: Claude 3 Haiku Anthropic · Speed & cost	72	—	—	$0.25 / $1.25	200K	96.0	Mar 2024
175	Mistral: Mixtral 8x7B InstructOSS Mistral AI · Open-source	72	—	—	$0.54 / $0.54	33K	133.3	Dec 2023
176	Xiaomi: MiMo-V2-Omni · Speed & cost	66	—	—	$0.4 / $2	262K	55.0	Mar 2026
177	Xiaomi: MiMo-V2-Pro · Speed & cost	66	—	—	$1 / $3	1M	33.0	Mar 2026
178	Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview) Google · Image generation	66	—	—	$0.5 / $3	66K	37.7	Feb 2026
179	AionLabs: Aion-2.0 · Speed & cost	66	—	—	$0.8 / $1.6	131K	55.0	Feb 2026
180	Z.ai: GLM 5OSS · Open-source	66	—	—	$0.72 / $2.3	80K	43.7	Feb 2026
181	MoonshotAI: Kimi K2.5 · Speed & cost	66	—	—	$0.3827 / $1.72	262K	62.8	Jan 2026
182	Writer: Palmyra X5 · Speed & cost	66	—	—	$0.6 / $6	1M	20.0	Jan 2026
183	OpenAI: GPT Audio Mini OpenAI · Speed & cost	66	—	—	$0.6 / $2.4	128K	44.0	Jan 2026
184	Z.ai: GLM 4.7OSS · Open-source	66	—	—	$0.39 / $1.75	203K	61.7	Dec 2025
185	Mistral: Devstral 2 2512OSS Mistral AI · Open-source	66	—	—	$0.4 / $2	262K	55.0	Dec 2025
186	Relace: Relace Search · Search + citations	66	—	—	$1 / $3	256K	33.0	Dec 2025
187	MoonshotAI: Kimi K2 Thinking · Hard reasoning	66	—	—	$0.47 / $2	131K	53.4	Nov 2025
188	Z.ai: GLM 4.6OSS · Open-source	66	—	—	$0.39 / $1.9	205K	57.6	Sep 2025
189	Relace: Relace Apply 3 · Speed & cost	66	—	—	$0.85 / $1.25	256K	62.9	Sep 2025
190	MoonshotAI: Kimi K2 0905 · Speed & cost	66	—	—	$0.4 / $2	131K	55.0	Sep 2025
191	Nous: Hermes 4 405BOSS · Search + citations	66	—	—	$1 / $3	131K	33.0	Aug 2025
192	Mistral: Mistral Medium 3.1OSS Mistral AI · Open-source	66	—	—	$0.4 / $2	131K	55.0	Aug 2025
193	Z.ai: GLM 4.5VOSS · Open-source	66	—	—	$0.6 / $1.8	66K	55.0	Aug 2025
194	Z.ai: GLM 4.5OSS · Open-source	66	—	—	$0.6 / $2.2	131K	47.1	Jul 2025
195	Switchpoint Router · Speed & cost	66	—	—	$0.85 / $3.4	131K	31.1	Jul 2025
196	MoonshotAI: Kimi K2 0711 · Speed & cost	66	—	—	$0.57 / $2.3	131K	46.0	Jul 2025
197	Mistral: Devstral MediumOSS Mistral AI · Open-source	66	—	—	$0.4 / $2	131K	55.0	Jul 2025
198	Morph: Morph V3 Large · Speed & cost	66	—	—	$0.9 / $1.9	262K	47.1	Jul 2025
199	Morph: Morph V3 Fast · Speed & cost	66	—	—	$0.8 / $1.2	82K	66.0	Jul 2025
200	Baidu: ERNIE 4.5 VL 424B A47B · Speed & cost	66	—	—	$0.42 / $1.25	123K	79.0	Jun 2025
201	MiniMax: MiniMax M1 · Speed & cost	66	—	—	$0.4 / $2.2	1M	50.8	Jun 2025
202	Mistral: Mistral Medium 3OSS Mistral AI · Open-source	66	—	—	$0.4 / $2	131K	55.0	May 2025
203	Arcee AI: Maestro Reasoning · Speed & cost	66	—	—	$0.9 / $3.3	131K	31.4	May 2025
204	Arcee AI: Virtuoso Large · Speed & cost	66	—	—	$0.75 / $1.2	131K	67.7	May 2025
205	Arcee AI: Coder Large · Code generation	66	—	—	$0.5 / $0.8	33K	101.5	May 2025
206	EleutherAI: Llemma 7b · Speed & cost	66	—	—	$0.8 / $1.2	4K	66.0	Apr 2025
207	AlfredPros: CodeLLaMa 7B Instruct SolidityOSS · Open-source	66	—	—	$0.8 / $1.2	4K	66.0	Apr 2025
208	NVIDIA: Llama 3.1 Nemotron Ultra 253B v1OSS · Open-source	66	—	—	$0.6 / $1.8	131K	55.0	Apr 2025
209	TheDrummer: Skyfall 36B V2 · Speed & cost	66	—	—	$0.55 / $0.8	33K	97.8	Mar 2025
210	AionLabs: Aion-1.0-Mini · Speed & cost	66	—	—	$0.7 / $1.4	131K	62.9	Feb 2025
211	Qwen: Qwen VL MaxOSS Alibaba Cloud · Open-source	66	—	—	$0.52 / $2.08	131K	50.8	Feb 2025
212	Qwen: Qwen2.5 VL 72B InstructOSS Alibaba Cloud · Open-source	66	—	—	$0.8 / $0.8	33K	82.5	Feb 2025
213	Perplexity: Sonar Perplexity · Search + citations	66	—	—	$1 / $1	127K	66.0	Jan 2025
214	Sao10K: Llama 3.3 Euryale 70BOSS · Hard reasoning	66	—	—	$0.65 / $0.75	131K	94.3	Dec 2024
215	Amazon: Nova Pro 1.0 Amazon · Speed & cost	66	—	—	$0.8 / $3.2	300K	33.0	Dec 2024
216	Qwen2.5 Coder 32B InstructOSS Alibaba Cloud · Code generation	66	—	—	$0.66 / $1	33K	79.5	Nov 2024
217	TheDrummer: UnslopNemo 12B · Speed & cost	66	—	—	$0.4 / $0.4	33K	165.0	Nov 2024
218	Sao10K: Llama 3.1 Euryale 70B v2.2OSS · Hard reasoning	66	—	—	$0.85 / $0.85	131K	77.6	Aug 2024
219	Nous: Hermes 3 405B InstructOSS · Search + citations	66	—	—	$1 / $1	131K	66.0	Aug 2024
220	Meta: Llama 3 70B InstructOSS Meta · Open-source	66	—	—	$0.51 / $0.74	8K	105.6	Apr 2024
221	OpenAI: GPT-3.5 Turbo (older v0613) OpenAI · Speed & cost	66	—	—	$1 / $2	4K	44.0	Jan 2024
222	Mancer: Weaver (alpha) · Speed & cost	66	—	—	$0.75 / $1	8K	75.4	Aug 2023
223	ReMM SLERP 13B · Speed & cost	66	—	—	$0.45 / $0.65	6K	120.0	Jul 2023
224	OpenAI: GPT-3.5 Turbo OpenAI · Speed & cost	66	—	—	$0.5 / $1.5	16K	66.0	May 2023
225	Google: Gemma 3 4BOSS Google · Open-source	65	—	—	$0.04 / $0.08	131K	1083.3	Mar 2025
226	AionLabs: Aion-RP 1.0 (8B)OSS · Open-source	65	—	—	$0.8 / $1.6	33K	54.2	Feb 2025
227	Microsoft: Phi 4OSS · Speed & cost	65	—	—	$0.065 / $0.14	16K	634.1	Jan 2025
228	Meta: Llama 3.1 8B InstructOSS Meta · Open-source	65	—	—	$0.02 / $0.05	16K	1857.1	Jul 2024
229	Google: Gemma 2 27BOSS Google · Open-source	65	—	—	$0.65 / $0.65	8K	100.0	Jul 2024
230	Google: Gemma 2 9BOSS Google · Open-source	65	—	—	$0.03 / $0.09	8K	1083.3	Jun 2024
231	NousResearch: Hermes 2 Pro - Llama-3 8BOSS · Search + citations	65	—	—	$0.14 / $0.14	8K	464.3	May 2024
232	Meta: Llama 3 8B InstructOSS Meta · Open-source	65	—	—	$0.03 / $0.04	8K	1857.1	Apr 2024
233	Google: Gemini 3.1 Flash Lite Preview Google · Speed & cost	62	—	—	$0.25 / $1.5	1M	70.9	Mar 2026
234	NVIDIA: Nemotron 3 Nano 30B A3B · Speed & cost	62	—	—	$0.05 / $0.2	262K	496.0	Dec 2025
235	NVIDIA: Nemotron Nano 12B 2 VL · Speed & cost	62	—	—	$0.2 / $0.6	131K	155.0	Oct 2025
236	IBM: Granite 4.0 Micro · Speed & cost	62	—	—	$0.017 / $0.11	131K	976.4	Oct 2025
237	NVIDIA: Nemotron Nano 9B V2 · Speed & cost	62	—	—	$0.04 / $0.16	131K	620.0	Sep 2025
238	Amazon: Nova Micro 1.0 Amazon · Speed & cost	62	—	—	$0.035 / $0.14	128K	708.6	Dec 2024
239	WizardLM-2 8x22BOSS · Speed & cost	62	—	—	$0.62 / $0.62	66K	100.0	Apr 2024
240	Arcee AI: Trinity Large Thinking · Hard reasoning	58	—	—	$0.22 / $0.85	262K	108.4	Apr 2026
241	Kwaipilot: KAT-Coder-Pro V2 · Code generation	58	—	—	$0.3 / $1.2	256K	77.3	Mar 2026
242	Reka Edge · Speed & cost	58	—	—	$0.1 / $0.1	16K	580.0	Mar 2026
243	MiniMax: MiniMax M2.7 · Speed & cost	58	—	—	$0.3 / $1.2	205K	77.3	Mar 2026
244	NVIDIA: Nemotron 3 Super · Speed & cost	58	—	—	$0.1 / $0.5	262K	193.3	Mar 2026
245	ByteDance Seed: Seed-2.0-Lite · Speed & cost	58	—	—	$0.25 / $2	262K	51.6	Mar 2026
246	Inception: Mercury 2 · Speed & cost	58	—	—	$0.25 / $0.75	128K	116.0	Mar 2026
247	ByteDance Seed: Seed-2.0-Mini · Speed & cost	58	—	—	$0.1 / $0.4	262K	232.0	Feb 2026
248	MiniMax: MiniMax M2.5 · Speed & cost	58	—	—	$0.118 / $0.99	197K	104.7	Feb 2026
249	StepFun: Step 3.5 Flash · Speed & cost	58	—	—	$0.1 / $0.3	262K	290.0	Jan 2026
250	Upstage: Solar Pro 3 · Speed & cost	58	—	—	$0.15 / $0.6	128K	154.7	Jan 2026
251	MiniMax: MiniMax M2-her · Speed & cost	58	—	—	$0.3 / $1.2	66K	77.3	Jan 2026
252	Z.ai: GLM 4.7 FlashOSS · Speed & cost	58	—	—	$0.06 / $0.4	203K	252.2	Jan 2026
253	AllenAI: Olmo 3.1 32B InstructOSS · Open-source	58	—	—	$0.2 / $0.6	66K	145.0	Jan 2026
254	ByteDance Seed: Seed 1.6 Flash · Speed & cost	58	—	—	$0.075 / $0.3	262K	309.3	Dec 2025
255	ByteDance Seed: Seed 1.6 · Speed & cost	58	—	—	$0.25 / $2	262K	51.6	Dec 2025
256	MiniMax: MiniMax M2.1 · Speed & cost	58	—	—	$0.27 / $0.95	197K	95.1	Dec 2025
257	Xiaomi: MiMo-V2-Flash · Speed & cost	58	—	—	$0.09 / $0.29	262K	305.3	Dec 2025
258	Z.ai: GLM 4.6VOSS · Open-source	58	—	—	$0.3 / $0.9	131K	96.7	Dec 2025
259	EssentialAI: Rnj 1 Instruct · Speed & cost	58	—	—	$0.15 / $0.15	33K	386.7	Dec 2025
260	Amazon: Nova 2 Lite Amazon · Speed & cost	58	—	—	$0.3 / $2.5	1M	41.4	Dec 2025
261	Mistral: Ministral 3 14B 2512OSS Mistral AI · Speed & cost	58	—	—	$0.2 / $0.2	262K	290.0	Dec 2025
262	Mistral: Ministral 3 8B 2512OSS Mistral AI · Speed & cost	58	—	—	$0.15 / $0.15	262K	386.7	Dec 2025
263	Mistral: Ministral 3 3B 2512OSS Mistral AI · Speed & cost	58	—	—	$0.1 / $0.1	131K	580.0	Dec 2025
264	Prime Intellect: INTELLECT-3 · Speed & cost	58	—	—	$0.2 / $1.1	131K	89.2	Nov 2025
265	AllenAI: Olmo 3 32B ThinkOSS · Hard reasoning	58	—	—	$0.15 / $0.5	66K	178.5	Nov 2025
266	xAI: Grok 4.1 Fast xAI · Speed & cost	58	—	—	$0.2 / $0.5	2M	165.7	Nov 2025
267	OpenAI: GPT-5.1-Codex-Mini OpenAI · Code generation	58	—	—	$0.25 / $2	400K	51.6	Nov 2025
268	Mistral: Voxtral Small 24B 2507OSS Mistral AI · Open-source	58	—	—	$0.1 / $0.3	32K	290.0	Oct 2025
269	OpenAI: gpt-oss-safeguard-20b OpenAI · Speed & cost	58	—	—	$0.075 / $0.3	131K	309.3	Oct 2025
270	MiniMax: MiniMax M2 · Speed & cost	58	—	—	$0.255 / $1	197K	92.4	Oct 2025
271	Baidu: ERNIE 4.5 21B A3B Thinking · Hard reasoning	58	—	—	$0.07 / $0.28	131K	331.4	Oct 2025
272	TheDrummer: Cydonia 24B V4.1 · Speed & cost	58	—	—	$0.3 / $0.5	131K	145.0	Sep 2025
273	xAI: Grok 4 Fast xAI · Speed & cost	58	—	—	$0.2 / $0.5	2M	165.7	Sep 2025
274	Tongyi DeepResearch 30B A3BOSS Alibaba Cloud · Search + citations	58	—	—	$0.09 / $0.45	131K	214.8	Sep 2025
275	Meituan: LongCat Flash Chat · Speed & cost	58	—	—	$0.2 / $0.8	131K	116.0	Sep 2025
276	Qwen: Qwen Plus 0728 (thinking)OSS Alibaba Cloud · Hard reasoning	58	—	—	$0.26 / $0.78	1M	111.5	Sep 2025
277	Qwen: Qwen Plus 0728OSS Alibaba Cloud · Open-source	58	—	—	$0.26 / $0.78	1M	111.5	Sep 2025
278	xAI: Grok Code Fast 1 xAI · Speed & cost	58	—	—	$0.2 / $1.5	256K	68.2	Aug 2025
279	Nous: Hermes 4 70BOSS · Search + citations	58	—	—	$0.13 / $0.4	131K	218.9	Aug 2025
280	Baidu: ERNIE 4.5 21B A3B · Speed & cost	58	—	—	$0.07 / $0.28	120K	331.4	Aug 2025
281	Baidu: ERNIE 4.5 VL 28B A3B · Speed & cost	58	—	—	$0.14 / $0.56	30K	165.7	Aug 2025
282	Z.ai: GLM 4.5 AirOSS · Open-source	58	—	—	$0.13 / $0.85	131K	118.4	Jul 2025
283	Z.ai: GLM 4 32B OSS · Open-source	58	—	—	$0.1 / $0.1	128K	580.0	Jul 2025
284	ByteDance: UI-TARS 7B · Speed & cost	58	—	—	$0.1 / $0.2	128K	386.7	Jul 2025
285	Mistral: Devstral Small 1.1OSS Mistral AI · Open-source	58	—	—	$0.1 / $0.3	131K	290.0	Jul 2025
286	Tencent: Hunyuan A13B Instruct · Speed & cost	58	—	—	$0.14 / $0.57	131K	163.4	Jul 2025
287	Baidu: ERNIE 4.5 300B A47B · Speed & cost	58	—	—	$0.28 / $1.1	123K	84.1	Jun 2025
288	Inception: Mercury · Speed & cost	58	—	—	$0.25 / $0.75	128K	116.0	Jun 2025
289	Arcee AI: Spotlight · Speed & cost	58	—	—	$0.18 / $0.18	131K	322.2	May 2025
290	Inception: Mercury Coder · Code generation	58	—	—	$0.25 / $0.75	128K	116.0	Apr 2025
291	Meta: Llama Guard 4 12BOSS Meta · Open-source	58	—	—	$0.18 / $0.18	164K	322.2	Apr 2025
292	Qwen: Qwen2.5 VL 32B InstructOSS Alibaba Cloud · Open-source	58	—	—	$0.2 / $0.6	128K	145.0	Mar 2025
293	Reka Flash 3 · Speed & cost	58	—	—	$0.1 / $0.2	66K	386.7	Mar 2025
294	Qwen: QwQ 32BOSS Alibaba Cloud · Hard reasoning	58	—	—	$0.15 / $0.58	131K	158.9	Mar 2025
295	Mistral: SabaOSS Mistral AI · Open-source	58	—	—	$0.2 / $0.6	33K	145.0	Feb 2025
296	Qwen: Qwen VL PlusOSS Alibaba Cloud · Open-source	58	—	—	$0.1365 / $0.4095	131K	212.5	Feb 2025
297	Qwen: Qwen-PlusOSS Alibaba Cloud · Open-source	58	—	—	$0.26 / $0.78	1M	111.5	Feb 2025
298	MiniMax: MiniMax-01 · Speed & cost	58	—	—	$0.2 / $1.1	1M	89.2	Jan 2025
299	Amazon: Nova Lite 1.0 Amazon · Speed & cost	58	—	—	$0.06 / $0.24	300K	386.7	Dec 2024
300	TheDrummer: Rocinante 12B · Speed & cost	58	—	—	$0.17 / $0.43	33K	193.3	Sep 2024
301	Meta: Llama 3.2 3B InstructOSS Meta · Open-source	58	—	—	$0.051 / $0.34	80K	296.7	Sep 2024
302	Cohere: Command R (08-2024)OSS Cohere · Open-source	58	—	—	$0.15 / $0.6	128K	154.7	Aug 2024
303	Mistral: Mistral 7B Instruct v0.1OSS Mistral AI · Open-source	58	—	—	$0.11 / $0.19	3K	386.7	Sep 2023
304	MythoMax 13B · Speed & cost	58	—	—	$0.06 / $0.06	4K	966.7	Jul 2023
305	LiquidAI: LFM2-24B-A2B · Speed & cost	50	—	—	$0.03 / $0.12	33K	666.7	Feb 2026
306	Arcee AI: Trinity Mini · Speed & cost	50	—	—	$0.045 / $0.15	131K	512.8	Dec 2025
307	OpenAI: gpt-oss-120b OpenAI · Speed & cost	50	—	—	$0.039 / $0.19	131K	436.7	Aug 2025
308	OpenAI: gpt-oss-20b OpenAI · Speed & cost	50	—	—	$0.03 / $0.11	131K	714.3	Aug 2025
309	Google: Gemma 3n 4BOSS Google · Open-source	50	—	—	$0.02 / $0.04	33K	1666.7	May 2025
310	Qwen: Qwen2.5 Coder 7B InstructOSS Alibaba Cloud · Code generation	50	—	—	$0.03 / $0.09	33K	833.3	Apr 2025
311	AllenAI: Olmo 2 32B InstructOSS · Open-source	50	—	—	$0.05 / $0.2	128K	400.0	Mar 2025
312	Llama Guard 3 8BOSS Meta · Open-source	50	—	—	$0.02 / $0.06	131K	1250.0	Feb 2025
313	Qwen: Qwen-TurboOSS Alibaba Cloud · Open-source	50	—	—	$0.0325 / $0.13	131K	615.4	Feb 2025
314	Cohere: Command R7B (12-2024)OSS Cohere · Open-source	50	—	—	$0.0375 / $0.15	128K	533.3	Dec 2024
315	Qwen: Qwen2.5 7B InstructOSS Alibaba Cloud · Open-source	50	—	—	$0.04 / $0.1	33K	714.3	Oct 2024
316	Meta: Llama 3.2 1B InstructOSS Meta · Open-source	50	—	—	$0.027 / $0.2	60K	440.5	Sep 2024
317	Meta: Llama 3.2 11B Vision InstructOSS Meta · Open-source	50	—	—	$0.049 / $0.049	131K	1020.4	Sep 2024
318	Sao10K: Llama 3 8B LunarisOSS · Hard reasoning	50	—	—	$0.04 / $0.05	8K	1111.1	Aug 2024

Quality = composite benchmark (MMLU, HumanEval, MATH)Arena ELO = LMSYS Chatbot Arena ratingValue = quality per dollarPrice = input / output per 1M tokens

What to do this quarter

Treat Arena Elo as a triage filter, not a decision. Use it to drop the bottom half of your candidate list, then run a real eval on the remainder.
Pick the right Arena board. Coding teams should read the coding Arena (Claude Opus 4.7 leads at 1567 Elo). Long-context teams should read the hard-prompts Arena. The aggregate text leaderboard is the wrong signal for many enterprise workloads.
Discount short-conversation polish. The Arena rewards style. Models tuned for chat win at the margin against models tuned for accuracy. Build internal evals that reward what your business actually pays for.
Watch the gap, not the ranking. Sub-25 Elo shifts are within statistical noise. Anything under 50 Elo between two candidates is a coin flip on most workloads.
Plan for the four-way race. Gemini 3.1 Pro, Claude Opus 4.7, GPT-5.5 Pro, and DeepSeek V4 Pro are approximately interchangeable on quality at the top. Optimise your stack for switching cost, not for capability.
Capture vote-rate momentum. The fastest-rising models week-over-week are usually the next month's leaders. Subscribe to weekly Arena reports.
Pair Arena Elo with cost. A 50-Elo lead at 10x the price is rarely a good trade. See our model leaderboard for combined quality-cost rankings.

LMArena Leaderboard — May 2026

What is the LMArena, in one paragraph?

How to read an Arena Elo score

Live Leaderboard

What to do this quarter

Related reading