Benchmarks / Arena (formerly LMArena)

Reported by Arena (formerly LMArena)

Arena (formerly LMArena)

Human preference on prompts that set explicit instructions.

Last updated 2 Oct 2026

Results dated
2 Oct 2026
Results
413 configurations of 387 models
Unit
Arena rating
Licence
Creative Commons Attribution 4.0 International

Arena (formerly LMArena) in brief

What it measures. Arena (formerly LMArena) is a third-party benchmark, shown here with credit. Measured here: Human preference on prompts that set explicit instructions.

Latest results. , 413 configurations ranked.

Top 3 on Instruction following (Arena rating, higher is better):

  1. Gemini 4 Argon (high reasoning) (Google): 1,528
  2. Claude Opus 5.5 (high reasoning) (Anthropic): 1,517
  3. Claude Opus 4.6 (high reasoning) (Anthropic): 1,514

The first 4 are too close to tell apart, so read the order among them as a tie.

How to cite. Arena (formerly LMArena), via Spring Prompt, results dated 2 October 2026. https://springprompt.com/benchmarks/lmarena These results belong to Arena (formerly LMArena), not to us. Licence: Creative Commons Attribution 4.0 International. Credit Arena (formerly LMArena) first.

Instruction following: Mistral Large (2402)

Top 15 of 413 results · Arena rating, higher is better · lines show the 95% range · ≈ cannot be told apart from the leader. Choose a model to highlight it.Clear highlight

  1. 1≈ Gemini 4 Argon (high reasoning)Google 1,528
  2. 2≈ Claude Opus 5.5 (high reasoning)Anthropic 1,517
  3. 3≈ Claude Opus 4.6 (high reasoning)Anthropic 1,514
  4. 4≈ Claude Fable 5 (high reasoning)Anthropic 1,509
  5. 5 Claude Opus 4.7 (high reasoning)Anthropic 1,502
  6. 6 Claude Opus 4.6Anthropic 1,499
  7. 7 Claude Fable 5.1 (max reasoning)Anthropic 1,497
  8. 8 Claude Opus 5 (high reasoning)Anthropic 1,496
  9. 9 Claude Opus 4.7Anthropic 1,493
  10. 9 Claude Opus 5 (max reasoning)Anthropic 1,493
  11. 11 Claude Opus 4.8 (high reasoning)Anthropic 1,491
  12. 12 GPT-5.6 Sol (extra-high reasoning)OpenAI 1,490
  13. 12 GPT-6.1 Sol (max reasoning)OpenAI 1,490
  14. 14 Kimi K3 (max reasoning)Moonshot AI 1,488
  15. 15 Gemini 3.8 Flash (high reasoning)Google 1,486
  16. 318 Mistral Large (2402)Mistral AI 1,238

Full results

Arena Text: instruction following, Arena rating, higher is better
#ModelInstruction following · 95% range
Arena rating, higher is better
Price
$ per million tokens, in / out
1≈ Gemini 4 Argon (high reasoning)Google
1,528
1,514–1,543
$2 / $10
2≈ Claude Opus 5.5 (high reasoning)Anthropic
1,517
1,502–1,533
$4 / $20
3≈ Claude Opus 4.6 (high reasoning)Anthropic
1,514
1,508–1,519
$5 / $25
4≈ Claude Fable 5 (high reasoning)Anthropic
1,509
1,503–1,515
$10 / $50
5 Claude Opus 4.7 (high reasoning)Anthropic
1,502
1,497–1,508
$5 / $25
6 Claude Opus 4.6Anthropic
1,499
1,494–1,504
$5 / $25
7 Claude Fable 5.1 (max reasoning)Anthropic
1,497
1,487–1,508
$10 / $50
8 Claude Opus 5 (high reasoning)Anthropic
1,496
1,490–1,502
$5 / $25
9 Claude Opus 4.7Anthropic
1,493
1,488–1,498
$5 / $25
9 Claude Opus 5 (max reasoning)Anthropic
1,493
1,486–1,500
$5 / $25
11 Claude Opus 4.8 (high reasoning)Anthropic
1,491
1,485–1,496
$5 / $25
12 GPT-5.6 Sol (extra-high reasoning)OpenAI
1,490
1,484–1,497
$4 / $20
12 GPT-6.1 Sol (max reasoning)OpenAI
1,490
1,471–1,509
$2 / $10
14 Kimi K3 (max reasoning)Moonshot AI
1,488
1,481–1,495
$3 / $15
15 Gemini 3.8 Flash (high reasoning)Google
1,486
1,479–1,493
$1.50 / $7.50
15 Muse Spark 1.3 (max reasoning)Meta
1,486
1,476–1,495
$1.25 / $4.25
17 Gemini 3.7 Flash (high reasoning)Google
1,485
1,477–1,492
$1.50 / $7.50
18 Claude Sonnet 5.5 (extra-high reasoning)Anthropic
1,484
1,466–1,502
$2 / $10
19 Claude Opus 4.5 (high reasoning, 32k budget)Anthropic
1,483
1,476–1,490
$5 / $25
20 Gemini 3.1 Pro PreviewGoogle
1,480
1,475–1,484
$2 / $12
21 DeepSeek-V4.1-Flash (max reasoning)DeepSeek
1,479
1,469–1,490
$0.30 / $1.20
22 Claude Opus 4.8Anthropic
1,478
1,473–1,484
$5 / $25
22 GPT-5.5 (high reasoning)OpenAI
1,478
1,472–1,483
$5 / $30
24 Muse Spark 1.2 (extra-high reasoning)Meta
1,477
1,462–1,492
$1.25 / $4.25
25 GLM-5.3 (max reasoning)Z.ai
1,476
1,467–1,484
$1.40 / $4.40
26 GPT-5.5OpenAI
1,475
1,470–1,480
$5 / $30
26 Claude Opus 4.5Anthropic
1,475
1,470–1,480
$5 / $25
28 Claude Sonnet 4.6Anthropic
1,474
1,469–1,480
$3 / $15
28 MiMo-V2.6-ProXiaomi
1,474
1,458–1,490
$0.43 / $0.87
28 GPT-6 Astra (max reasoning)OpenAI
1,474
1,463–1,485
$10 / $50
28 Qwen3.8-Max (0902)Alibaba
1,474
1,466–1,481
$2 / $6
32 Gemini 3 Pro PreviewGoogle
1,473
1,466–1,479
–
33 Muse Spark 1.1Meta
1,472
1,466–1,479
$1.25 / $4.25
33 Gemini 3.6 Flash (high reasoning)Google
1,472
1,466–1,479
$1.50 / $7.50
33 GLM-5.3-FlashZ.ai
1,472
1,464–1,479
$0.15 / $0.50
36 GPT-5.4 (high reasoning)OpenAI
1,471
1,466–1,477
$2.50 / $15
37 GLM-5.2 (max reasoning)Z.ai
1,470
1,464–1,475
$1.40 / $4.40
38 MiMo-V2.5-ProXiaomi
1,469
1,464–1,474
$0.43 / $0.87
39 Claude Sonnet 5 (high reasoning)Anthropic
1,467
1,461–1,472
$2 / $10
39 Qwen3.5-Max-PreviewAlibaba
1,467
1,459–1,474
–
41 Gemini 3.5 Flash (medium reasoning)Google
1,466
1,460–1,471
$1.50 / $9
42 Qwen3.7-Max-PreviewAlibaba
1,465
1,449–1,482
–
42 Gemini 3.5 Flash (high reasoning)Google
1,465
1,459–1,471
$1.50 / $9
42 Muse SparkMeta
1,465
1,456–1,474
–
45 Claude Sonnet 4.5 (high reasoning, 32k budget)Anthropic
1,464
1,459–1,468
$3 / $15
46 Claude Sonnet 4.5Anthropic
1,462
1,457–1,466
$3 / $15
47 GPT-5.6 Terra (extra-high reasoning)OpenAI
1,461
1,455–1,468
$2 / $12
47 DeepSeek-V4-Pro (0813, high reasoning)DeepSeek
1,461
1,451–1,471
$1.32 / $3.96
47 Grok 4.5xAI
1,461
1,455–1,467
$2 / $6
50 GLM-5.1Z.ai
1,460
1,455–1,465
$1.33 / $4.18
50 GPT-5.4OpenAI
1,460
1,455–1,465
$2.50 / $15
52 Claude Opus 4.1 (16k reasoning budget)Anthropic
1,458
1,452–1,464
$15 / $75
53 Gemini 3 Flash PreviewGoogle
1,457
1,450–1,465
$0.50 / $3
53 GPT-5.2 Chat (2026-02-10)OpenAI
1,457
1,450–1,463
–
55 GPT-5.5 InstantOpenAI
1,456
1,449–1,463
–
55 Kimi K2.6Moonshot AI
1,456
1,449–1,462
$0.95 / $4
57 Claude Opus 4.1Anthropic
1,454
1,450–1,459
$15 / $75
58 MiMo-V2.6-FlashXiaomi
1,453
1,440–1,466
$0.14 / $0.28
58 Gemma 4 31BGoogle
1,453
1,439–1,467
$0.15 / $0.40
58 GPT-6 Sol (max reasoning)OpenAI
1,453
1,442–1,464
$2 / $10
58 ERNIE 5.1Baidu
1,453
1,446–1,459
–
62 DeepSeek-V4-Pro (0423)DeepSeek
1,452
1,447–1,458
$1.42 / $2.83
63 Grok 4.6 (high reasoning)xAI
1,451
1,443–1,458
$2 / $6
64 GPT-5.1 (high reasoning)OpenAI
1,450
1,444–1,457
$1.25 / $10
64 Grok 4.20 Beta 1xAI
1,450
1,443–1,457
–
64 Qwen3.6-Max-PreviewAlibaba
1,450
1,436–1,463
–
67 DeepSeek-V4-Pro (0423, high reasoning)DeepSeek
1,448
1,442–1,454
$1.42 / $2.83
67 GPT-6 Luna (max reasoning)OpenAI
1,448
1,438–1,458
$0.10 / $0.50
69 Qwen3.7-PlusAlibaba
1,447
1,441–1,453
$0.32 / $1.28
69 Grok 4.20 Beta (0309, reasoning)xAI
1,447
1,441–1,452
–
71 GLM-5Z.ai
1,446
1,440–1,453
$1 / $3.20
71 Hy3Tencent
1,446
1,436–1,456
$0.14 / $0.58
73 Step 5 PreviewStepFun
1,445
1,427–1,464
$1 / $2.70
73 GPT-5.6 Luna (extra-high reasoning)OpenAI
1,445
1,439–1,451
$0.20 / $1.20
73 MiMo-V2-ProXiaomi
1,445
1,437–1,452
–
76 Claude Opus 4 (16k reasoning budget)Anthropic
1,444
1,437–1,451
–
76 Grok 4.20 Multi-Agent Beta (0309)xAI
1,444
1,439–1,449
–
78 Gemini 3.5 Flash-LiteGoogle
1,443
1,437–1,449
$0.30 / $2.50
78 Gemini 3 Flash Preview (minimal reasoning)Google
1,443
1,438–1,448
$0.50 / $3
80 Grok 4.7 (extra-high reasoning)xAI
1,440
1,427–1,453
$2 / $6
81 Kimi K2.5 (reasoning on)Moonshot AI
1,439
1,434–1,444
$0.57 / $2.85
82 Gemini 2.5 ProGoogle
1,438
1,434–1,442
$1.25 / $10
83 Gemma 4 26B A4BGoogle
1,437
1,424–1,451
$0.09 / $0.30
83 GPT-4.5 PreviewOpenAI
1,437
1,428–1,445
–
85 GPT-5.3 ChatOpenAI
1,436
1,430–1,443
–
86 Qwen3.6-PlusAlibaba
1,435
1,430–1,441
$0.33 / $1.95
86 Dola-Seed-2.0-ProByteDance
1,435
1,430–1,440
–
88 MiniMax-M3MiniMax
1,434
1,429–1,440
$0.30 / $1.20
88 GPT-5.4 mini (high reasoning)OpenAI
1,434
1,428–1,439
$0.75 / $4.50
88 Qwen3.5-397B-A17BAlibaba
1,434
1,429–1,438
$0.55 / $3.50
91 DeepSeek-V4-Flash (0423, high reasoning)DeepSeek
1,433
1,428–1,439
$0.05 / $0.45
92 Kimi K2.5 (no reasoning)Moonshot AI
1,432
1,420–1,444
$0.57 / $2.85
93 Grok 4.1xAI
1,431
1,426–1,436
–
93 MiMo-V2.5Xiaomi
1,431
1,424–1,437
$0.40 / $2
95 Grok 4.1 ThinkingxAI
1,430
1,424–1,435
–
96 DeepSeek-V4-Flash (0423)DeepSeek
1,428
1,422–1,434
$0.05 / $0.45
96 MiMo-V2-OmniXiaomi
1,428
1,419–1,436
–
98 GPT-5.2 (high reasoning)OpenAI
1,427
1,421–1,433
$1.75 / $14
98 ChatGPT-4o (2025-03-26)OpenAI
1,427
1,422–1,431
–
100 GPT-5.1OpenAI
1,426
1,421–1,432
$1.25 / $10
100 Qwen3.8-27BAlibaba
1,426
1,418–1,434
$0.50 / $3
100 InklingThinking Machines
1,426
1,420–1,432
$1 / $4.05
103 GLM-4.7Z.ai
1,425
1,415–1,436
$0.60 / $2.20
103 ERNIE 5.0 (0110)Baidu
1,425
1,419–1,431
–
105 Qwen3-Max-PreviewAlibaba
1,424
1,417–1,432
–
105 GPT-5.2OpenAI
1,424
1,419–1,429
$1.75 / $14
105 GLM-5V-TurboZ.ai
1,424
1,414–1,434
$1.20 / $4
108 Mistral Medium 3.5Mistral AI
1,422
1,412–1,432
$1.50 / $7.50
109 DeepSeek-V3.1-Terminus (reasoning on)DeepSeek
1,420
1,400–1,440
$0.27 / $1
109 DeepSeek-V3.2DeepSeek
1,420
1,414–1,426
$0.28 / $0.42
109 DeepSeek-V3.2 (reasoning on)DeepSeek
1,420
1,414–1,425
$0.28 / $0.42
112 ERNIE 5.0 Preview (1203)Baidu
1,419
1,408–1,431
–
113 Kimi K2 Thinking TurboMoonshot AI
1,418
1,413–1,423
–
114 DeepSeek-V3.1 (reasoning on)DeepSeek
1,416
1,405–1,427
$0.27 / $1
114 DeepSeek-V3.2-Exp (reasoning on)DeepSeek
1,416
1,404–1,427
$0.27 / $0.41
114 LongCat-Flash-Chat (2602, experimental)Meituan
1,416
1,409–1,423
–
117 Claude Sonnet 4 (32k reasoning budget)Anthropic
1,415
1,409–1,422
–
117 Claude Haiku 4.5Anthropic
1,415
1,411–1,419
$1 / $5
117 Claude Opus 4Anthropic
1,415
1,409–1,422
–
117 GPT-5 ChatOpenAI
1,415
1,408–1,422
–
121 DeepSeek-V3.2-ExpDeepSeek
1,414
1,404–1,425
$0.27 / $0.41
121 Grok 4.3xAI
1,414
1,409–1,420
$1.25 / $2.50
121 Qwen3-235B-A22B-Instruct-2507Alibaba
1,414
1,410–1,418
$0.15 / $0.75
121 Nova Experimental Chat (2026-02-10)Amazon
1,414
1,396–1,432
–
125 Qwen3-MaxAlibaba
1,413
1,402–1,425
–
125 Muse GlimmerMeta
1,413
1,397–1,429
–
127 GLM-4.6Z.ai
1,412
1,406–1,419
$0.50 / $2
127 Claude 3.7 Sonnet (32k reasoning budget)Anthropic
1,412
1,405–1,418
–
129 GPT-5 (high reasoning)OpenAI
1,410
1,403–1,417
$1.25 / $10
130 Qwen3-VL-235B-A22B-InstructAlibaba
1,409
1,398–1,421
$0.21 / $1.90
131 Gemini 3.1 Flash-Lite PreviewGoogle
1,407
1,402–1,413
$0.25 / $1.50
131 o1OpenAI
1,407
1,401–1,414
–
131 MiniMax-M2.7MiniMax
1,407
1,402–1,412
$0.30 / $1.20
134 Qwen3.5-122B-A10BAlibaba
1,406
1,399–1,412
$0.26 / $2.08
135 Nemotron 3 Ultra 550B A55B (NVFP4)NVIDIA
1,405
1,395–1,415
–
135 GLM-4.5Z.ai
1,405
1,397–1,412
$0.60 / $2.20
137 GPT-4.1OpenAI
1,403
1,398–1,409
$2 / $8
137 Grok 3 Preview (02-24)xAI
1,403
1,397–1,410
–
137 o3OpenAI
1,403
1,397–1,408
$2 / $8
140 DeepSeek-V3.1DeepSeek
1,402
1,392–1,412
$0.27 / $1
140 Grok 4.1 Fast (reasoning)xAI
1,402
1,396–1,407
–
142 Qwen3.5-27BAlibaba
1,401
1,394–1,407
$0.27 / $2.16
143 Mistral Large 3Mistral AI
1,400
1,396–1,405
$0.50 / $1.50
143 Gemini 2.5 Flash Preview (09-2025)Google
1,400
1,393–1,407
–
145 Gemini 2.5 FlashGoogle
1,399
1,396–1,403
$0.30 / $2.50
146 Nova Experimental Chat (2026-01-10)Amazon
1,398
1,380–1,416
–
146 ERNIE 5.0 Preview (1022)Baidu
1,398
1,383–1,413
–
148 DeepSeek-R1DeepSeek
1,397
1,390–1,405
–
148 Hunyuan Vision 1.5 ThinkingTencent
1,397
1,375–1,419
–
148 Grok 4 (0709)xAI
1,397
1,390–1,403
–
151 Hy3 PreviewTencent
1,396
1,384–1,409
$0.18 / $0.60
151 Claude Sonnet 4Anthropic
1,396
1,389–1,403
–
153 Grok 4 Fast ChatxAI
1,395
1,381–1,410
–
153 Mistral Medium 3.1Mistral AI
1,395
1,391–1,399
$0.40 / $2
155 DeepSeek-V3.1-TerminusDeepSeek
1,394
1,377–1,412
$0.27 / $1
155 Inkling SmallThinking Machines
1,394
1,387–1,400
$0.45 / $1.20
157 DeepSeek-R1-0528DeepSeek
1,391
1,381–1,401
$0.50 / $2.18
158 Kimi K2 (0905)Moonshot AI
1,390
1,379–1,401
$0.60 / $2.50
159 Grok 4 Fast (reasoning)xAI
1,388
1,380–1,397
–
159 LongCat-Flash-ChatMeituan
1,388
1,377–1,399
–
161 GPT-5.4 nano (high reasoning)OpenAI
1,387
1,381–1,392
$0.20 / $1.25
162 Qwen3.5-35B-A3BAlibaba
1,386
1,379–1,392
$0.15 / $1
163 Claude 3.7 SonnetAnthropic
1,385
1,380–1,391
–
163 MiniMax-M2.1 PreviewMiniMax
1,385
1,376–1,394
–
163 Qwen3-235B-A22B-Thinking-2507Alibaba
1,385
1,372–1,397
$0.45 / $3.50
166 Step 3.5 FlashStepFun
1,384
1,379–1,389
$0.10 / $0.30
166 Qwen3-Coder-480B-A35BAlibaba
1,384
1,376–1,391
$0.30 / $1
168 Qwen3-VL-235B-A22B-ThinkingAlibaba
1,382
1,370–1,394
–
168 MiMo-V2-Flash (no reasoning)Xiaomi
1,382
1,377–1,388
–
170 MiniMax-M2.5MiniMax
1,381
1,375–1,387
$0.29 / $1.20
170 Qwen3.5-FlashAlibaba
1,381
1,376–1,386
$0.065 / $0.26
170 o1-previewOpenAI
1,381
1,374–1,388
–
170 Qwen3-235B-A22B (no reasoning)Alibaba
1,381
1,374–1,388
–
170 Kimi K2 (0711)Moonshot AI
1,381
1,373–1,388
$0.57 / $2.30
175 Solar Pro 4Upstage
1,379
1,371–1,387
$0.15 / $0.60
176 DeepSeek-V3-0324DeepSeek
1,378
1,372–1,384
$0.25 / $1
177 Nova Experimental Chat (12-10)Amazon
1,377
1,359–1,396
–
178 Qwen3-Next-80B-A3B-InstructAlibaba
1,376
1,369–1,384
$0.09 / $1.10
178 MiMo-V2-Flash (reasoning on)Xiaomi
1,376
1,365–1,386
–
180 Claude 3.5 Sonnet (2024-10-22)Anthropic
1,374
1,370–1,379
–
181 Hunyuan T1 (2025-07-11)Tencent
1,373
1,356–1,391
–
181 GPT-5 mini (high reasoning)OpenAI
1,373
1,366–1,381
$0.25 / $2
181 GPT-4.1 miniOpenAI
1,373
1,366–1,379
$0.40 / $1.60
184 Trinity Large PreviewArcee AI
1,370
1,363–1,376
–
185 o4-miniOpenAI
1,369
1,362–1,375
–
186 Gemini 2.5 Flash-Lite Preview (06-17, reasoning on)Google
1,368
1,361–1,375
–
187 GLM-4.6VZ.ai
1,367
1,346–1,388
$0.30 / $0.90
187 Qwen3-30B-A3B-Instruct-2507Alibaba
1,367
1,359–1,374
$0.09 / $0.30
189 o3-mini (high reasoning)OpenAI
1,366
1,359–1,374
–
189 Mistral Medium 3Mistral AI
1,366
1,359–1,373
$0.40 / $2
191 Gemini 2.5 Flash-Lite Preview (09-2025, no reasoning)Google
1,363
1,358–1,369
–
192 Nova Experimental Chat (11-10)Amazon
1,361
1,354–1,368
–
192 GLM-4.5-AirZ.ai
1,361
1,354–1,368
$0.14 / $0.86
194 Grok 3 Mini (high reasoning)xAI
1,360
1,351–1,369
–
195 Qwen3-Next-80B-A3B-ThinkingAlibaba
1,359
1,350–1,369
–
196 Trinity Large ThinkingArcee AI
1,357
1,350–1,364
–
196 Qwen3-235B-A22BAlibaba
1,357
1,349–1,364
–
196 Qwen2.5-MaxAlibaba
1,357
1,351–1,363
–
199 Hunyuan TurboS (2025-04-16)Tencent
1,353
1,341–1,365
–
199 Grok 3 Mini BetaxAI
1,353
1,345–1,361
–
201 Llama 3.1 Nemotron Ultra 253B v1NVIDIA
1,351
1,330–1,373
–
201 Hunyuan TurboS (2025-02-26)Tencent
1,351
1,333–1,368
–
203 GLM-4.7-FlashZ.ai
1,348
1,338–1,358
$0.06 / $0.40
204 Gemini 2.0 Flash (001)Google
1,347
1,342–1,353
–
205 MiniMax-M1MiniMax
1,346
1,339–1,353
–
206 Nova Experimental Chat (10-20)Amazon
1,345
1,335–1,356
–
206 Step 3StepFun
1,345
1,330–1,359
–
208 Nemotron 3.5 Lightning 30B A3B (NVFP4)NVIDIA
1,344
1,335–1,353
–
208 o3-miniOpenAI
1,344
1,339–1,349
–
210 DeepSeek-V3DeepSeek
1,343
1,336–1,350
$0.26 / $1.03
210 Gemma 3 27BGoogle
1,343
1,337–1,349
$0.08 / $0.35
212 Nemotron 3 Super 120B A12BNVIDIA
1,342
1,330–1,355
$0.08 / $0.45
212 Command ACohere
1,342
1,336–1,347
$2.50 / $10
214 Gemini 1.5 Pro (002)Google
1,340
1,335–1,345
–
215 GLM-4.5VZ.ai
1,339
1,323–1,355
$0.60 / $1.80
216 Claude 3.5 Sonnet (2024-06-20)Anthropic
1,338
1,333–1,343
–
216 Granite 4.2 30BIBM
1,338
1,320–1,356
–
216 Mistral Small 3.2 24BMistral AI
1,338
1,329–1,347
$0.094 / $0.25
219 MiniMax-M2MiniMax
1,334
1,321–1,347
–
220 INTELLECT-3Prime Intellect
1,332
1,317–1,348
–
220 Qwen-PlusAlibaba
1,332
1,320–1,344
$0.26 / $0.78
220 o1-miniOpenAI
1,332
1,326–1,337
–
223 Gemini 2.0 Flash-Lite Preview (02-05)Google
1,331
1,325–1,338
–
223 Qwen3-32BAlibaba
1,331
1,312–1,350
$0.14 / $0.40
225 Nova 2 LiteAmazon
1,327
1,317–1,337
$0.30 / $2.50
225 GPT-5 nano (high reasoning)OpenAI
1,327
1,314–1,340
$0.05 / $0.40
227 Mercury 2Inception
1,326
1,307–1,346
–
227 GPT-4o (2024-05-13)OpenAI
1,326
1,321–1,331
$5 / $15
229 Llama 3.3 Nemotron Super 49B v1NVIDIA
1,325
1,306–1,345
–
230 gpt-oss-120bOpenAI
1,323
1,316–1,330
$0.10 / $0.50
230 QwQ-32BAlibaba
1,323
1,316–1,330
–
232 Llama 3.3 Nemotron Super 49B v1.5NVIDIA
1,321
1,301–1,341
–
232 Gemma 3 12BGoogle
1,321
1,305–1,337
$0.05 / $0.15
234 Olmo 3.1 32B InstructAi2
1,320
1,310–1,331
–
235 GPT-4o (2024-08-06)OpenAI
1,319
1,313–1,325
$2.50 / $10
236 Gemini Advanced (0514)Google
1,317
1,311–1,324
–
237 Ring-flash-2.0Ant Group
1,316
1,303–1,330
–
237 Claude 3.5 HaikuAnthropic
1,316
1,311–1,321
–
237 Hunyuan Large (2025-02-10)Tencent
1,316
1,300–1,332
–
237 Step-2 16K Exp (2024-12)StepFun
1,316
1,303–1,328
–
241 Ling-flash-2.0Ant Group
1,315
1,302–1,329
–
241 Llama 3.1 405B Instruct (FP8)Meta
1,315
1,310–1,320
–
243 Claude 3 OpusAnthropic
1,314
1,310–1,319
–
243 GLM-4-Plus (0111)Z.ai
1,314
1,302–1,327
–
243 Nova Experimental Chat (10-09)Amazon
1,314
1,293–1,336
–
243 Llama 3.1 405B Instruct (BF16)Meta
1,314
1,309–1,320
–
243 DeepSeek-V2.5-1210DeepSeek
1,314
1,303–1,326
–
243 Llama 4 MaverickMeta
1,314
1,307–1,320
$0.35 / $1
243 Hunyuan Turbo (0110)Tencent
1,314
1,295–1,332
–
250 Grok 2 (2024-08-13)xAI
1,312
1,307–1,317
–
250 Gemini 1.5 Pro (001)Google
1,312
1,306–1,317
–
252 Qwen3-30B-A3BAlibaba
1,310
1,302–1,317
$0.12 / $0.50
253 Yi-Lightning01.AI
1,308
1,301–1,315
–
254 Step-1o Turbo (2025-06)StepFun
1,307
1,294–1,319
–
255 Molmo 2 8BAi2
1,304
1,267–1,342
–
256 GPT-4 TurboOpenAI
1,303
1,298–1,309
–
257 Athene-V2-ChatNexusflow
1,302
1,296–1,308
–
257 Qwen-Max (0919)Alibaba
1,302
1,294–1,310
–
259 GLM-4-PlusZ.ai
1,301
1,294–1,308
–
259 Magistral Medium 1.0Mistral AI
1,301
1,290–1,311
–
261 GPT-4.1 nanoOpenAI
1,300
1,288–1,313
$0.10 / $0.40
261 Mistral Large 2 (2407)Mistral AI
1,300
1,294–1,305
$2 / $6
263 Llama 4 ScoutMeta
1,299
1,292–1,306
$0.10 / $0.30
263 Olmo 3 32B ThinkAi2
1,299
1,283–1,315
–
265 GPT-4 Turbo Preview (1106)OpenAI
1,297
1,291–1,302
–
266 Mistral Large 2.1 (2411)Mistral AI
1,295
1,289–1,301
–
267 Qwen2.5-Plus (1127)Alibaba
1,294
1,285–1,304
–
267 Mistral Small 3.1 24BMistral AI
1,294
1,287–1,301
$0.35 / $0.56
269 GPT-4o mini (2024-07-18)OpenAI
1,293
1,288–1,298
$0.15 / $0.60
269 Llama 3.3 70B InstructMeta
1,293
1,288–1,298
$0.71 / $0.71
271 Qwen2.5-72B-InstructAlibaba
1,292
1,286–1,298
$0.36 / $0.40
271 DeepSeek-V2.5DeepSeek
1,292
1,285–1,298
–
273 Nemotron 3 Nano 30B A3BNVIDIA
1,291
1,282–1,300
$0.06 / $0.24
274 GPT-4 Turbo Preview (0125)OpenAI
1,290
1,285–1,296
–
274 Gemini 1.5 Flash (002)Google
1,290
1,284–1,296
–
276 Granite 4.1 8BIBM
1,287
1,270–1,304
–
277 Hunyuan Large VisionTencent
1,286
1,269–1,303
–
278 GPT-4 (0314)OpenAI
1,285
1,278–1,293
–
278 Granite 4.2 3BIBM
1,285
1,265–1,305
–
280 Grok 2 Mini (2024-08-13)xAI
1,284
1,278–1,289
–
281 Gemma 3n E4BGoogle
1,282
1,273–1,290
–
281 Hunyuan Standard (2025-02-10)Tencent
1,282
1,266–1,297
–
283 gpt-oss-20bOpenAI
1,281
1,269–1,294
$0.03 / $0.14
283 Granite 4.2 8BIBM
1,281
1,262–1,300
$0.06 / $0.15
285 Llama 3.1 Nemotron 70B InstructNVIDIA
1,280
1,269–1,291
–
285 Olmo 3.1 32B ThinkAi2
1,280
1,266–1,293
–
287 Athene 70B (0725)Nexusflow
1,279
1,271–1,286
–
288 GPT-4 (0613)OpenAI
1,277
1,271–1,283
–
288 Nova Pro 1.0Amazon
1,277
1,270–1,283
$0.80 / $3.20
290 Llama 3.1 70B InstructMeta
1,273
1,267–1,278
$0.72 / $0.72
291 Llama 3.1 Tülu 3 70BAi2
1,272
1,257–1,288
–
292 Gemma 2 27BGoogle
1,271
1,267–1,276
$0.65 / $0.65
293 MercuryInception
1,269
1,244–1,295
–
294 Claude 3 SonnetAnthropic
1,268
1,263–1,274
–
294 Granite H SmallIBM
1,268
1,252–1,283
–
296 Gemma 3 4BGoogle
1,267
1,251–1,284
$0.04 / $0.05
296 Jamba 1.5 LargeAI21 Labs
1,267
1,256–1,277
–
298 Gemini 1.5 Flash (001)Google
1,265
1,259–1,271
–
299 Qwen2.5-Coder-32B-InstructAlibaba
1,264
1,252–1,276
$0.66 / $1
300 Llama 3.1 Nemotron 51B InstructNVIDIA
1,263
1,248–1,277
–
301 Reka Core (2024-09-04)Reka AI
1,262
1,252–1,272
–
302 Nemotron-4 340B InstructNVIDIA
1,259
1,251–1,268
–
302 Llama 3 70B InstructMeta
1,259
1,253–1,264
–
304 GLM-4 (0520)Z.ai
1,256
1,246–1,267
–
305 Mistral Small 3Mistral AI
1,255
1,247–1,264
$0.05 / $0.08
306 Command R+ (08-2024)Cohere
1,254
1,245–1,263
$2.50 / $10
307 DeepSeek-Coder-V2DeepSeek
1,253
1,244–1,262
–
308 Gemma 2 9B SimPOPrinceton NLP
1,252
1,242–1,262
–
309 Reka Flash (2024-09-04)Reka AI
1,249
1,239–1,260
–
309 Aya Expanse 32BCohere
1,249
1,243–1,256
–
311 Phi-4Microsoft
1,245
1,239–1,252
$0.07 / $0.14
311 Claude 3 HaikuAnthropic
1,245
1,240–1,251
–
311 Gemma 2 9BGoogle
1,245
1,240–1,251
–
314 Nova Lite 1.0Amazon
1,244
1,237–1,251
$0.06 / $0.24
315 Hunyuan Standard 256KTencent
1,243
1,227–1,260
–
315 Qwen2-72B-InstructAlibaba
1,243
1,236–1,249
–
317 Command R+Cohere
1,241
1,235–1,247
–
318 Gemini 1.5 Flash-8B (001)Google
1,238
1,232–1,244
–
318 Mistral Large (2402)Mistral AI
1,238
1,231–1,245
–
320 Command R (08-2024)Cohere
1,236
1,227–1,245
$0.15 / $0.60
321 OLMo 2 32B Instruct (0325)Ai2
1,229
1,212–1,246
–
322 Gemini ProGoogle
1,220
1,204–1,237
–
323 Qwen1.5-110B-ChatAlibaba
1,218
1,210–1,226
–
324 GPT-3.5 Turbo (0125)OpenAI
1,216
1,209–1,222
–
325 Mixtral 8x22B InstructMistral AI
1,215
1,209–1,222
$2 / $6
325 Nova Micro 1.0Amazon
1,215
1,208–1,222
$0.035 / $0.14
327 Qwen1.5-72B-ChatAlibaba
1,212
1,205–1,220
–
328 Ministral 8B (2410)Mistral AI
1,211
1,198–1,224
–
328 Mistral MediumMistral AI
1,211
1,202–1,219
–
330 Gemini Pro (Dev API)Google
1,210
1,200–1,221
–
331 Llama 3.1 Tülu 3 8BAi2
1,208
1,193–1,224
–
332 Jamba 1.5 MiniAI21 Labs
1,205
1,194–1,215
–
332 Aya Expanse 8BCohere
1,205
1,195–1,214
–
334 Reka Flash 21B (2024-02-26, online)Reka AI
1,202
1,192–1,212
–
335 Command RCohere
1,200
1,193–1,207
–
336 GPT-3.5 Turbo (1106)OpenAI
1,198
1,186–1,210
–
337 Zephyr ORPO 141B A35B v0.1Hugging Face
1,194
1,178–1,209
–
338 Granite 3.1 8B InstructIBM
1,193
1,176–1,210
–
338 Reka Flash 21B (2024-02-26)Reka AI
1,193
1,184–1,201
–
338 Llama 3 8B InstructMeta
1,193
1,187–1,199
–
341 Llama 3.1 8B InstructMeta
1,191
1,185–1,197
$0.05 / $0.08
342 Yi-1.5-34B-Chat01.AI
1,188
1,180–1,196
–
342 DBRX Instruct PreviewDatabricks
1,188
1,179–1,196
–
344 Qwen1.5-32B-ChatAlibaba
1,185
1,176–1,193
–
345 Mixtral 8x7B Instruct v0.1Mistral AI
1,181
1,175–1,187
–
346 InternLM2.5 20B ChatShanghai AI Lab
1,178
1,168–1,188
–
346 Phi-3-medium-4k-instructMicrosoft
1,178
1,170–1,185
–
348 Granite 3.1 2B InstructIBM
1,173
1,157–1,190
–
349 Granite 3.0 8B InstructIBM
1,172
1,160–1,185
–
350 Tülu 2 DPO 70BAi2
1,171
1,157–1,186
–
350 Gemma 2 2BGoogle
1,171
1,165–1,177
–
352 Qwen1.5-14B-ChatAlibaba
1,168
1,158–1,178
–
353 WizardLM 70BMicrosoft
1,164
1,150–1,177
–
354 DeepSeek LLM 67B ChatDeepSeek
1,158
1,141–1,175
–
355 Gemma 1.1 7BGoogle
1,157
1,149–1,165
–
356 OpenChat 3.5 (0106)OpenChat
1,156
1,145–1,166
–
357 OpenChat 3.5OpenChat
1,155
1,140–1,169
–
358 Phi-3-small-8k-instructMicrosoft
1,154
1,146–1,163
–
358 Snowflake Arctic InstructSnowflake
1,154
1,145–1,162
–
358 OpenHermes 2.5 Mistral 7BTeknium
1,154
1,138–1,169
–
361 Yi-34B-Chat01.AI
1,152
1,143–1,162
–
362 Llama 3.2 3B InstructMeta
1,146
1,135–1,157
$0.05 / $0.33
362 Starling-LM-7B-betaNexusflow
1,146
1,136–1,156
–
362 QwQ-32B-PreviewAlibaba
1,146
1,130–1,161
–
365 Vicuna 33BLMSYS
1,140
1,131–1,149
–
366 Starling-LM-7B-alphaUC Berkeley
1,137
1,125–1,148
–
367 MPT-30B-ChatMosaicML
1,135
1,114–1,156
–
367 Llama 2 70B ChatMeta
1,135
1,127–1,143
–
367 Falcon 180B ChatTII
1,135
1,106–1,163
–
370 Granite 3.0 2B InstructIBM
1,132
1,120–1,145
–
371 Dolphin 2.2.1 Mistral 7BCognitive Computations
1,129
1,105–1,154
–
372 Llama 2 70B SteerLM ChatNVIDIA
1,125
1,106–1,143
–
372 Phi-3-mini-4k-instruct (June 2024)Microsoft
1,125
1,115–1,134
–
374 WizardLM 13BMicrosoft
1,124
1,110–1,138
–
374 Qwen1.5-7B-ChatAlibaba
1,124
1,110–1,138
–
374 Mistral 7B Instruct v0.2Mistral AI
1,124
1,114–1,133
–
377 Qwen-14B-ChatAlibaba
1,119
1,103–1,135
–
378 PaLM 2Google
1,117
1,104–1,131
–
378 Vicuna 13BLMSYS
1,117
1,108–1,127
–
380 Solar 10.7B Instruct v1.0Upstage
1,116
1,098–1,135
–
381 Phi-3-mini-4k-instructMicrosoft
1,114
1,105–1,122
–
382 Llama 2 13B ChatMeta
1,110
1,101–1,120
–
383 Nous Hermes 2 Mixtral 8x7B DPONous Research
1,108
1,091–1,124
–
384 Gemma 7BGoogle
1,105
1,092–1,118
–
384 Code Llama 34B InstructMeta
1,105
1,091–1,118
–
386 SmolLM2 1.7B InstructHugging Face
1,104
1,083–1,125
–
387 Gemma 1.1 2BGoogle
1,102
1,091–1,113
–
388 Phi-3-mini-128k-instructMicrosoft
1,100
1,090–1,111
–
389 Zephyr 7B AlphaHugging Face
1,099
1,075–1,123
–
390 Code Llama 70B InstructMeta
1,098
1,068–1,127
–
391 StripedHyena Nous 7BTogether AI
1,091
1,076–1,106
–
392 Zephyr 7B BetaHugging Face
1,089
1,076–1,102
–
393 Mistral 7B InstructMistral AI
1,087
1,074–1,101
–
394 Llama 3.2 1B InstructMeta
1,086
1,075–1,097
$0.027 / $0.20
395 Vicuna 7BLMSYS
1,077
1,063–1,091
–
396 Gemma 2BGoogle
1,071
1,055–1,087
–
397 Qwen1.5-4B-ChatAlibaba
1,070
1,057–1,083
–
397 Llama 2 7B ChatMeta
1,070
1,060–1,080
–
399 Guanaco 33BTim Dettmers
1,066
1,045–1,087
–
400 GPT4All 13B SnoozyNomic AI
1,038
1,013–1,063
–
401 ChatGLM3-6BZ.ai
1,037
1,019–1,055
–
402 OLMo 7B InstructAi2
1,029
1,013–1,045
–
403 Koala 13BUC Berkeley
1,027
1,011–1,042
–
404 Alpaca 13BStanford
1,025
1,009–1,042
–
405 MPT-7B-ChatMosaicML
1,009
991–1,028
–
406 OpenAssistant Pythia 12BOpenAssistant
988
972–1,004
–
407 ChatGLM2-6BZ.ai
983
960–1,005
–
408 ChatGLM-6BZ.ai
978
960–996
–
409 RWKV-4 Raven 14BRWKV
970
953–987
–
410 FastChat-T5 3BLMSYS
958
940–977
–
411 Dolly v2 12BDatabricks
940
919–961
–
412 LLaMA 13BMeta
918
893–944
–
413 StableLM Tuned Alpha 7BStability AI
910
890–930
–

Swipe the table sideways for more columns.

Source: Arena (formerly LMArena), via Spring Prompt, springprompt.com/benchmarks/lmarena. Test your own prompt on these models →

Ranks follow the score as shown, so equal numbers share a rank. ≈ marks results whose 95% range overlaps the leader's: they cannot be told apart from it. Results as published by Arena (formerly LMArena); we do not re-run them.

Progress over time

Instruction following, plotted at each model's release date, higher is better. Each model counts once, at its best setting. Scores are today's, not launch day's; 178 models without a known release date are left off. Hover or tap a mark for its model.

Best score so far

Every model at its release date. The line moves each time a new release beats every earlier one, and each new best is named where there is room.

1,0001,2001,4001,600 2024Jul2025Jul2026Jul GLM-4 (0520)GPT-4 TurboGPT-4o (2024-05-13)Claude 3.5 Sonnet (2024-06-20)o1-previewo1Claude 3.7 SonnetGPT-4.5 PreviewClaude Opus 4Claude Opus 4.1Gemini 3 Pro PreviewClaude Opus 4.5Claude Opus 4.6Claude Opus 5.5Gemini 4 Argon

Best score by lab

Each lab's best so far, stepping whenever one of its releases beats its own earlier models. The 6 labs with the best scores today; 17 others not drawn.

GoogleAnthropicOpenAIMoonshot AIMetaDeepSeek

1,2001,3001,4001,5001,600 2024Jul2025Jul2026Jul GoogleAnthropicOpenAIMoonshot AIMetaDeepSeek

What it measures

Human preference on prompts that set explicit instructions.

What it does not measure

Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.

Contains data from the Arena Leaderboard Dataset by Arena, licensed under CC BY 4.0. Licence: Creative Commons Attribution 4.0 International.