1,520
1,496–1,544
Benchmarks / Arena (formerly LMArena)
Reported by Arena (formerly LMArena)
Arena (formerly LMArena)
Human preference on writing and language prompts.
- Results dated
- 25 Sep 2026
- Models
- 408
- Unit
- Arena rating
Full results
| # | Model | Arena Text: writing, literature and language · 95% range Arena rating, higher is better |
|---|---|---|
| 1 | Claude Opus 5.5 (high)Anthropic | |
| 1 | Claude Fable 5 (high)Anthropic |
1,508 1,501–1,515 |
| 1 | Claude Opus 4.6 (high)Anthropic |
1,500 1,495–1,506 |
| 1 | Claude Opus 4.7 (high)Anthropic |
1,496 1,490–1,502 |
| 1 | Claude Fable 5.1 (max)Anthropic |
1,496 1,484–1,508 |
| 1 | Gemini 3.7 Flash (high)Google |
1,496 1,487–1,505 |
| 2 | Claude Opus 4.6Anthropic |
1,490 1,485–1,496 |
| 4 | Gemini 3.8 Flash (high)Google |
1,486 1,477–1,495 |
| 4 | Claude Opus 4.7Anthropic |
1,485 1,478–1,491 |
| 4 | Muse Spark 1.2Meta |
1,472 1,452–1,492 |
| 5 | GPT-5.6 Sol (xhigh)OpenAI |
1,483 1,475–1,490 |
| 5 | Claude Opus 5 (high)Anthropic |
1,482 1,476–1,489 |
| 5 | gemini-3-proGoogle |
1,481 1,474–1,488 |
| 5 | Claude Opus 5 (max)Anthropic |
1,480 1,473–1,488 |
| 6 | Gemini 3.1 Pro PreviewGoogle |
1,480 1,475–1,485 |
| 8 | Kimi K3 (max)Moonshot AI |
1,475 1,467–1,483 |
| 8 | Claude Opus 4.8 (high)Anthropic |
1,474 1,468–1,480 |
| 8 | Gemini 3.6 Flash (high)Google |
1,472 1,465–1,480 |
| 8 | Gemini 3.5 Flash (high)Google |
1,472 1,466–1,479 |
| 8 | Qwen3.8 Max (0902)Alibaba |
1,472 1,463–1,480 |
| 8 | GLM 5.3 (max)Z.ai |
1,470 1,460–1,480 |
| 8 | GPT-6 Astra (max)OpenAI |
1,469 1,455–1,483 |
| 9 | Muse Spark 1.3 (max)Meta |
1,466 1,454–1,478 |
| 9 | qwen3.7-max-previewAlibaba |
1,458 1,438–1,478 |
| 11 | GPT-5.5 (high)OpenAI |
1,470 1,464–1,475 |
| 13 | Gemini 3.5 Flash (medium)Google |
1,468 1,462–1,475 |
| 13 | MiMo-V2.6-ProXiaomi |
1,456 1,437–1,475 |
| 14 | Claude Opus 4.8Anthropic |
1,467 1,461–1,473 |
| 14 | Claude Opus 4.5Anthropic |
1,466 1,459–1,473 |
| 14 | muse-sparkMeta |
1,463 1,452–1,473 |
| 15 | GPT-5.5OpenAI |
1,467 1,461–1,472 |
| 15 | Claude Opus 4.5Anthropic |
1,467 1,461–1,472 |
| 15 | GPT-5.4 (high)OpenAI |
1,465 1,459–1,471 |
| 15 | gemini-3-flashGoogle |
1,461 1,453–1,469 |
| 15 | gpt-5.5-instantOpenAI |
1,461 1,453–1,469 |
| 15 | DeepSeek V4.1 Flash (max)DeepSeek |
1,455 1,440–1,470 |
| 16 | Muse Spark 1.1Meta |
1,460 1,453–1,467 |
| 17 | Grok 4.5xAI |
1,459 1,452–1,466 |
| 17 | DeepSeek V4 Pro 0813DeepSeek |
1,454 1,442–1,466 |
| 18 | GLM 5.2 (max)Z.ai |
1,459 1,452–1,466 |
| 19 | GPT-5.4OpenAI |
1,458 1,452–1,464 |
| 19 | qwen3.5-max-previewAlibaba |
1,455 1,447–1,464 |
| 19 | Qwen3.6 Max PreviewAlibaba |
1,446 1,429–1,464 |
| 25 | Claude Sonnet 4.6Anthropic |
1,455 1,449–1,461 |
| 25 | grok-4.20-beta1xAI |
1,453 1,445–1,461 |
| 26 | gpt-5.2-chat-latest-20260210OpenAI |
1,453 1,446–1,460 |
| 26 | GLM 5.1Z.ai |
1,453 1,447–1,459 |
| 27 | Claude Sonnet 4.5Anthropic |
1,454 1,449–1,459 |
| 27 | Claude Sonnet 5 (high)Anthropic |
1,452 1,446–1,459 |
| 28 | Claude Sonnet 4.5Anthropic |
1,452 1,447–1,457 |
| 28 | DeepSeek V4 Pro 0423DeepSeek |
1,451 1,445–1,457 |
| 28 | MiMo-V2.5-ProXiaomi |
1,450 1,444–1,456 |
| 28 | GLM 5.3 FlashZ.ai |
1,450 1,441–1,459 |
| 28 | grok-4.20-beta-0309-reasoningxAI |
1,449 1,443–1,455 |
| 28 | gpt-4.5-preview-2025-02-27OpenAI |
1,448 1,438–1,457 |
| 28 | GLM 5Z.ai |
1,447 1,440–1,455 |
| 28 | GPT-6 Sol (max)OpenAI |
1,440 1,423–1,458 |
| 30 | gemini-3-flash (minimal reasoning)Google |
1,448 1,443–1,453 |
| 30 | Grok 4.6 (high)xAI |
1,446 1,437–1,454 |
| 35 | GPT-5.6 Terra (xhigh)OpenAI |
1,445 1,438–1,452 |
| 36 | Claude Opus 4.1Anthropic |
1,446 1,440–1,452 |
| 38 | deepseek-v4-pro-preview (high reasoning)DeepSeek |
1,443 1,437–1,449 |
| 38 | Gemini 3.5 Flash LiteGoogle |
1,442 1,435–1,450 |
| 38 | Gemma 4 31BGoogle |
1,435 1,420–1,450 |
| 39 | Gemini 2.5 ProGoogle |
1,445 1,440–1,449 |
| 39 | Claude Opus 4.1Anthropic |
1,444 1,439–1,449 |
| 39 | ernie-5.1Baidu |
1,442 1,435–1,449 |
| 40 | grok-4.20-multi-agent-beta-0309xAI |
1,443 1,437–1,449 |
| 40 | Kimi K2.6Moonshot AI |
1,441 1,434–1,448 |
| 48 | GPT-5.1 (high)OpenAI |
1,437 1,430–1,444 |
| 48 | Qwen3.7 PlusAlibaba |
1,437 1,430–1,443 |
| 48 | Grok 4.7 (xhigh)xAI |
1,426 1,407–1,444 |
| 51 | GPT-5.6 Luna (xhigh)OpenAI |
1,435 1,428–1,442 |
| 51 | Hy3Tencent |
1,429 1,417–1,441 |
| 56 | mimo-v2-proXiaomi |
1,431 1,423–1,439 |
| 57 | claude-opus-4 (thinking-16k reasoning)Anthropic |
1,431 1,424–1,438 |
| 60 | MiMo-V2.6-FlashXiaomi |
1,418 1,400–1,437 |
| 64 | chatgpt-4o-latest-20250326OpenAI |
1,431 1,426–1,436 |
| 64 | gpt-5.3-chat-latestOpenAI |
1,428 1,420–1,435 |
| 67 | Kimi K2.5Moonshot AI |
1,428 1,423–1,433 |
| 67 | grok-4.1-thinkingxAI |
1,427 1,422–1,433 |
| 67 | grok-4.1xAI |
1,427 1,421–1,433 |
| 67 | GPT-5.1OpenAI |
1,424 1,418–1,431 |
| 67 | GPT-6 Luna (max)OpenAI |
1,415 1,398–1,433 |
| 67 | DeepSeek V3.1 TerminusDeepSeek |
1,411 1,390–1,433 |
| 69 | Grok 4.3xAI |
1,424 1,418–1,430 |
| 69 | GPT-5.4 Mini (high)OpenAI |
1,424 1,418–1,430 |
| 69 | Qwen3.6 PlusAlibaba |
1,423 1,416–1,430 |
| 69 | deepseek-v4-flash-preview (high reasoning)DeepSeek |
1,422 1,416–1,429 |
| 70 | MiniMax M3MiniMax |
1,422 1,416–1,428 |
| 70 | DeepSeek V3.1 TerminusDeepSeek |
1,408 1,387–1,428 |
| 71 | claude-opus-4Anthropic |
1,420 1,413–1,426 |
| 71 | GLM 4.7Z.ai |
1,416 1,405–1,427 |
| 71 | ernie-5.0-preview-1203Baidu |
1,415 1,403–1,427 |
| 72 | Qwen3.5 397B A17BAlibaba |
1,421 1,416–1,426 |
| 72 | Gemini 3.1 Flash Lite PreviewGoogle |
1,420 1,414–1,426 |
| 72 | ernie-5.0-0110Baidu |
1,419 1,412–1,426 |
| 72 | GLM 5V TurboZ.ai |
1,412 1,400–1,424 |
| 76 | DeepSeek V4 Flash 0423DeepSeek |
1,416 1,409–1,422 |
| 77 | DeepSeek V3.2 ExpDeepSeek |
1,410 1,399–1,422 |
| 78 | Gemma 4 26B A4BGoogle |
1,406 1,391–1,421 |
| 80 | MiMo-V2.5Xiaomi |
1,412 1,405–1,418 |
| 80 | DeepSeek V3.1DeepSeek |
1,407 1,396–1,419 |
| 80 | ernie-5.0-preview-1022Baidu |
1,402 1,385–1,419 |
| 82 | Qwen3 MaxAlibaba |
1,405 1,392–1,418 |
| 83 | Kimi K2.5Moonshot AI |
1,404 1,391–1,417 |
| 84 | GPT-5.2OpenAI |
1,411 1,406–1,417 |
| 86 | DeepSeek V3.2DeepSeek |
1,410 1,403–1,416 |
| 88 | dola-seed-2.0-proBytedance |
1,409 1,403–1,415 |
| 88 | GPT-5.2 (high)OpenAI |
1,408 1,402–1,414 |
| 88 | Mistral Medium 3.5Mistral |
1,404 1,392–1,415 |
| 89 | grok-4-1-fast-reasoningxAI |
1,407 1,402–1,413 |
| 89 | gpt-5-chatOpenAI |
1,407 1,399–1,414 |
| 89 | grok-3-preview-02-24xAI |
1,406 1,399–1,413 |
| 89 | qwen3-max-previewAlibaba |
1,405 1,398–1,413 |
| 91 | GPT-4.1OpenAI |
1,405 1,399–1,412 |
| 91 | GLM 4.6Z.ai |
1,405 1,398–1,412 |
| 91 | DeepSeek V3.2DeepSeek |
1,403 1,397–1,410 |
| 91 | InklingThinkingmachines |
1,403 1,396–1,411 |
| 91 | mimo-v2-omniXiaomi |
1,403 1,393–1,412 |
| 91 | DeepSeek V3.2 ExpDeepSeek |
1,398 1,385–1,411 |
| 91 | grok-4-fast-chatxAI |
1,394 1,378–1,409 |
| 92 | DeepSeek V3.1DeepSeek |
1,398 1,388–1,408 |
| 93 | hunyuan-vision-1.5-thinkingTencent |
1,380 1,354–1,407 |
| 94 | Gemini 2.5 FlashGoogle |
1,402 1,398–1,406 |
| 94 | grok-4-0709xAI |
1,399 1,392–1,405 |
| 94 | o1OpenAI |
1,398 1,391–1,405 |
| 94 | Claude Sonnet 4Anthropic |
1,398 1,391–1,405 |
| 94 | Qwen3.8 27BAlibaba |
1,396 1,386–1,406 |
| 96 | Claude Haiku 4.5Anthropic |
1,399 1,395–1,404 |
| 96 | kimi-k2-thinking-turboMoonshot AI |
1,399 1,393–1,404 |
| 96 | GPT-5 (high)OpenAI |
1,397 1,389–1,404 |
| 96 | muse-glimmerMeta |
1,385 1,365–1,404 |
| 99 | o3OpenAI |
1,396 1,390–1,402 |
| 101 | Qwen3 235B A22B Instruct 2507Alibaba |
1,396 1,392–1,401 |
| 101 | gemini-2.5-flash-preview-09-2025Google |
1,394 1,387–1,402 |
| 101 | claude-3-7-sonnet-20250219 (thinking-32k reasoning)Anthropic |
1,394 1,387–1,401 |
| 101 | nvidia-nemotron-3-ultra-550b-a55b-nvfp4NVIDIA |
1,389 1,377–1,400 |
| 101 | Qwen3 VL 235B A22B InstructAlibaba |
1,388 1,375–1,401 |
| 102 | longcat-flash-chat-2602-expMeituan |
1,392 1,384–1,400 |
| 108 | R1 0528DeepSeek |
1,388 1,379–1,398 |
| 111 | Claude Sonnet 4Anthropic |
1,390 1,383–1,397 |
| 111 | grok-4-fast-reasoningxAI |
1,388 1,379–1,397 |
| 115 | Qwen3.5-122B-A10BAlibaba |
1,387 1,380–1,395 |
| 115 | Kimi K2 0905Moonshot AI |
1,382 1,370–1,394 |
| 115 | amazon-nova-experimental-chat-26-02-10Amazon |
1,374 1,353–1,394 |
| 116 | Mistral Large 3 2512Mistral |
1,388 1,383–1,393 |
| 117 | DeepSeek V3 0324DeepSeek |
1,386 1,380–1,393 |
| 117 | Hy3 previewTencent |
1,378 1,363–1,393 |
| 121 | MiniMax M2.7MiniMax |
1,386 1,380–1,391 |
| 124 | Mistral Medium 3.1Mistral |
1,386 1,382–1,391 |
| 124 | Qwen3 235B A22B Thinking 2507Alibaba |
1,377 1,363–1,391 |
| 125 | R1DeepSeek |
1,382 1,374–1,390 |
| 125 | GLM 4.5Z.ai |
1,382 1,374–1,391 |
| 126 | amazon-nova-experimental-chat-12-10Amazon |
1,371 1,351–1,390 |
| 127 | o1-previewOpenAI |
1,382 1,374–1,390 |
| 128 | Qwen3.5-27BAlibaba |
1,380 1,372–1,388 |
| 128 | amazon-nova-experimental-chat-26-01-10Amazon |
1,368 1,348–1,389 |
| 131 | gemini-2.5-flash-lite-preview-06-17 (thinking reasoning)Google |
1,379 1,372–1,387 |
| 131 | hunyuan-t1-20250711Tencent |
1,368 1,350–1,387 |
| 133 | claude-3-7-sonnet-20250219Anthropic |
1,379 1,373–1,386 |
| 133 | Qwen3 235B A22BAlibaba |
1,378 1,370–1,385 |
| 138 | claude-3-5-sonnet-20241022Anthropic |
1,378 1,373–1,383 |
| 138 | Kimi K2 0711Moonshot AI |
1,375 1,367–1,383 |
| 140 | GLM 4.6VZ.ai |
1,357 1,333–1,380 |
| 142 | Qwen3 VL 235B A22B ThinkingAlibaba |
1,365 1,351–1,380 |
| 145 | gemini-2.5-flash-lite-preview-09-2025 (no reasoning)Google |
1,371 1,365–1,378 |
| 145 | Mistral Medium 3Mistral |
1,370 1,363–1,378 |
| 146 | MiniMax M2.5MiniMax |
1,369 1,363–1,376 |
| 146 | Qwen3 Coder 480B A35BAlibaba |
1,366 1,358–1,375 |
| 149 | minimax-m2.1-previewMiniMax |
1,364 1,355–1,374 |
| 150 | Qwen3.5-FlashAlibaba |
1,367 1,361–1,373 |
| 152 | mimo-v2-flash (no reasoning)Xiaomi |
1,366 1,360–1,372 |
| 154 | Step 3.5 FlashStepfun |
1,366 1,360–1,372 |
| 154 | Qwen3.5-35B-A3BAlibaba |
1,364 1,356–1,371 |
| 156 | qwen2.5-maxAlibaba |
1,362 1,356–1,369 |
| 156 | GPT-5.4 Nano (high)OpenAI |
1,362 1,356–1,368 |
| 156 | GPT-4.1 MiniOpenAI |
1,361 1,354–1,368 |
| 156 | hunyuan-turbos-20250416Tencent |
1,357 1,345–1,368 |
| 156 | mimo-v2-flash (thinking reasoning)Xiaomi |
1,355 1,344–1,367 |
| 157 | trinity-large-previewArcee Ai |
1,359 1,351–1,366 |
| 157 | longcat-flash-chatMeituan |
1,354 1,342–1,366 |
| 159 | DeepSeek V3DeepSeek |
1,356 1,348–1,364 |
| 164 | Gemma 3 27BGoogle |
1,354 1,348–1,361 |
| 164 | gemini-1.5-pro-002Google |
1,354 1,348–1,360 |
| 164 | o4 MiniOpenAI |
1,354 1,348–1,361 |
| 166 | gemini-2.0-flash-001Google |
1,353 1,347–1,360 |
| 166 | GPT-5 Mini (high)OpenAI |
1,352 1,344–1,360 |
| 170 | gemini-2.0-flash-lite-preview-02-05Google |
1,347 1,340–1,355 |
| 172 | grok-3-mini (high reasoning)xAI |
1,344 1,334–1,354 |
| 172 | hunyuan-turbos-20250226Tencent |
1,333 1,312–1,354 |
| 173 | Command ACohere |
1,348 1,342–1,354 |
| 174 | Qwen3 235B A22BAlibaba |
1,345 1,337–1,353 |
| 174 | Inkling SmallThinkingmachines |
1,344 1,335–1,352 |
| 174 | Qwen3 Next 80B A3B ThinkingAlibaba |
1,341 1,330–1,352 |
| 175 | GPT-4o (2024-05-13)OpenAI |
1,345 1,340–1,351 |
| 176 | Trinity Large ThinkingArcee Ai |
1,343 1,336–1,351 |
| 177 | Qwen3 Next 80B A3B InstructAlibaba |
1,342 1,334–1,351 |
| 178 | Qwen3 30B A3B Instruct 2507Alibaba |
1,342 1,333–1,350 |
| 178 | o3 Mini HighOpenAI |
1,341 1,333–1,350 |
| 178 | GLM 4.5 AirZ.ai |
1,341 1,334–1,349 |
| 178 | grok-3-mini-betaxAI |
1,341 1,332–1,350 |
| 178 | gemini-advanced-0514Google |
1,341 1,333–1,348 |
| 178 | intellect-3Primeintellect |
1,333 1,317–1,350 |
| 180 | GPT-4o (2024-08-06)OpenAI |
1,341 1,334–1,348 |
| 184 | Solar Pro 4Upstage |
1,336 1,325–1,347 |
| 184 | glm-4-plus-0111Z.ai |
1,332 1,317–1,347 |
| 187 | GLM 4.7 FlashZ.ai |
1,331 1,320–1,343 |
| 187 | hunyuan-turbo-0110Tencent |
1,323 1,303–1,343 |
| 189 | gemini-1.5-pro-001Google |
1,334 1,327–1,341 |
| 189 | llama-3.1-nemotron-ultra-253b-v1NVIDIA |
1,321 1,300–1,341 |
| 191 | MiniMax M1MiniMax |
1,332 1,325–1,339 |
| 191 | Mistral Small 3.2 24BMistral |
1,329 1,319–1,339 |
| 191 | Nemotron 3 SuperNVIDIA |
1,325 1,311–1,338 |
| 192 | claude-3-5-sonnet-20240620Anthropic |
1,331 1,325–1,337 |
| 192 | GPT-4 TurboOpenAI |
1,330 1,324–1,337 |
| 192 | amazon-nova-experimental-chat-11-10Amazon |
1,328 1,320–1,336 |
| 192 | Qwen-PlusAlibaba |
1,322 1,307–1,336 |
| 194 | step-2-16k-exp-202412Stepfun |
1,319 1,304–1,335 |
| 194 | GLM 4.5VZ.ai |
1,318 1,300–1,335 |
| 198 | llama-3.3-nemotron-49b-super-v1NVIDIA |
1,311 1,288–1,334 |
| 199 | grok-2-2024-08-13xAI |
1,327 1,321–1,333 |
| 199 | step-3Stepfun |
1,318 1,302–1,333 |
| 203 | o3 MiniOpenAI |
1,325 1,319–1,331 |
| 203 | Gemma 3 12BGoogle |
1,313 1,295–1,331 |
| 204 | deepseek-v2.5-1210DeepSeek |
1,315 1,301–1,328 |
| 204 | hunyuan-large-2025-02-10Tencent |
1,310 1,291–1,329 |
| 204 | Mercury 2Inception |
1,307 1,285–1,329 |
| 204 | amazon-nova-experimental-chat-10-09Amazon |
1,305 1,282–1,329 |
| 205 | amazon-nova-experimental-chat-10-20Amazon |
1,315 1,304–1,327 |
| 208 | Qwen3 32BAlibaba |
1,308 1,291–1,325 |
| 209 | claude-3-opus-20240229Anthropic |
1,316 1,311–1,322 |
| 209 | llama-3.1-405b-instruct-fp8Meta |
1,316 1,310–1,322 |
| 212 | GPT-4o-mini (2024-07-18)OpenAI |
1,314 1,308–1,319 |
| 212 | nvidia-llama-3.3-nemotron-super-49b-v1.5NVIDIA |
1,297 1,276–1,319 |
| 213 | claude-3-5-haiku-20241022Anthropic |
1,314 1,309–1,319 |
| 214 | llama-3.1-405b-instruct-bf16Meta |
1,312 1,306–1,318 |
| 214 | Llama 4 MaverickMeta |
1,312 1,305–1,319 |
| 214 | glm-4-plusZ.ai |
1,309 1,301–1,318 |
| 214 | qwen-max-0919Alibaba |
1,307 1,298–1,317 |
| 215 | o1-miniOpenAI |
1,311 1,305–1,317 |
| 215 | gpt-oss-120bOpenAI |
1,309 1,301–1,317 |
| 216 | gemini-1.5-flash-002Google |
1,309 1,302–1,316 |
| 216 | gpt-4-1106-previewOpenAI |
1,308 1,301–1,314 |
| 216 | qwq-32bAlibaba |
1,308 1,300–1,315 |
| 216 | gpt-4-0125-previewOpenAI |
1,306 1,299–1,312 |
| 216 | Llama 4 ScoutMeta |
1,305 1,297–1,313 |
| 216 | Qwen3 30B A3BAlibaba |
1,304 1,296–1,312 |
| 216 | step-1o-turbo-202506Stepfun |
1,302 1,288–1,316 |
| 216 | olmo-3.1-32b-instructAi2 |
1,300 1,288–1,312 |
| 216 | GPT-4.1 NanoOpenAI |
1,299 1,285–1,314 |
| 216 | MiniMax M2MiniMax |
1,298 1,283–1,314 |
| 216 | Gemma 3 4BGoogle |
1,298 1,281–1,315 |
| 216 | granite-4.2-30bIBM |
1,294 1,273–1,316 |
| 217 | yi-lightning01 Ai |
1,303 1,295–1,311 |
| 217 | nvidia-nemotron-3.5-lightning-30b-a3b-nvfp4NVIDIA |
1,300 1,288–1,311 |
| 217 | hunyuan-standard-2025-02-10Tencent |
1,293 1,275–1,311 |
| 219 | Mistral Large 2407Mistral |
1,303 1,296–1,310 |
| 219 | Nova 2 LiteAmazon |
1,298 1,287–1,310 |
| 219 | hunyuan-large-visionTencent |
1,292 1,274–1,310 |
| 221 | gemma-3n-e4b-itGoogle |
1,300 1,291–1,309 |
| 224 | mistral-large-2411Mistral |
1,298 1,291–1,305 |
| 225 | ring-flash-2.0Ant Group |
1,289 1,274–1,305 |
| 228 | GPT-5 Nano (high)OpenAI |
1,289 1,275–1,304 |
| 229 | Gemma 2 27BGoogle |
1,298 1,292–1,303 |
| 229 | magistral-medium-2506Mistral |
1,291 1,278–1,303 |
| 235 | grok-2-mini-2024-08-13xAI |
1,295 1,288–1,301 |
| 235 | qwen2.5-plus-1127Alibaba |
1,289 1,278–1,300 |
| 239 | gemini-1.5-flash-001Google |
1,292 1,285–1,299 |
| 239 | Mistral Small 3.1 24BMistral |
1,290 1,283–1,298 |
| 241 | Llama 3.3 70B InstructMeta |
1,292 1,286–1,297 |
| 241 | llama-3.1-tulu-3-70bAi2 |
1,278 1,258–1,297 |
| 245 | gpt-4-0613OpenAI |
1,285 1,278–1,292 |
| 245 | deepseek-v2.5DeepSeek |
1,284 1,276–1,292 |
| 245 | gpt-4-0314OpenAI |
1,284 1,275–1,292 |
| 245 | ling-flash-2.0Ant Group |
1,277 1,262–1,293 |
| 245 | granite-4.1-8bIBM |
1,273 1,252–1,294 |
| 246 | olmo-3-32b-thinkAi2 |
1,275 1,258–1,291 |
| 250 | Qwen2.5 72B InstructAlibaba |
1,282 1,276–1,289 |
| 250 | jamba-1.5-largeAI21 Labs |
1,277 1,265–1,290 |
| 251 | gpt-oss-20bOpenAI |
1,275 1,262–1,288 |
| 252 | Command R+ (08-2024)Cohere |
1,277 1,266–1,288 |
| 255 | athene-70b-0725Nexusflow |
1,279 1,270–1,288 |
| 255 | llama-3.1-nemotron-51b-instructNVIDIA |
1,270 1,251–1,288 |
| 256 | athene-v2-chatNexusflow |
1,279 1,271–1,286 |
| 256 | gemma-2-9b-it-simpoPrinceton Nlp |
1,275 1,263–1,287 |
| 256 | mercuryInception |
1,257 1,228–1,286 |
| 257 | reka-core-20240904Rekaai |
1,272 1,258–1,286 |
| 260 | olmo-3.1-32b-thinkAi2 |
1,269 1,255–1,283 |
| 263 | Nova Pro 1.0Amazon |
1,274 1,267–1,281 |
| 264 | llama-3.1-nemotron-70b-instructNVIDIA |
1,265 1,251–1,279 |
| 264 | ibm-granite-h-smallIBM |
1,260 1,243–1,278 |
| 267 | claude-3-sonnet-20240229Anthropic |
1,270 1,263–1,276 |
| 267 | Nemotron 3 Nano 30B A3BNVIDIA |
1,267 1,256–1,277 |
| 273 | gemma-2-9b-itGoogle |
1,269 1,263–1,275 |
| 276 | reka-flash-20240904Rekaai |
1,259 1,246–1,272 |
| 277 | Llama 3.1 70B InstructMeta |
1,264 1,257–1,270 |
| 278 | command-r-plusCohere |
1,259 1,252–1,267 |
| 278 | nemotron-4-340b-instructNVIDIA |
1,258 1,249–1,268 |
| 279 | glm-4-0520Z.ai |
1,254 1,242–1,266 |
| 280 | Granite 4.2 8BIBM |
1,242 1,219–1,265 |
| 284 | c4ai-aya-expanse-32bCohere |
1,254 1,246–1,262 |
| 286 | gemini-1.5-flash-8b-001Google |
1,255 1,247–1,262 |
| 289 | Mistral Small 3Mistral |
1,248 1,238–1,258 |
| 289 | granite-4.2-3bIBM |
1,234 1,209–1,258 |
| 291 | llama-3-70b-instructMeta |
1,250 1,243–1,256 |
| 292 | claude-3-haiku-20240307Anthropic |
1,247 1,241–1,253 |
| 292 | deepseek-coder-v2DeepSeek |
1,243 1,232–1,253 |
| 292 | olmo-2-0325-32b-instructAi2 |
1,234 1,214–1,253 |
| 296 | qwen2-72b-instructAlibaba |
1,242 1,234–1,250 |
| 296 | Command R (08-2024)Cohere |
1,239 1,227–1,250 |
| 297 | Nova Lite 1.0Amazon |
1,240 1,231–1,248 |
| 301 | Qwen2.5 Coder 32B InstructAlibaba |
1,228 1,213–1,243 |
| 303 | ministral-8b-2410Mistral |
1,226 1,210–1,242 |
| 305 | Phi 4Microsoft |
1,229 1,222–1,237 |
| 306 | gpt-3.5-turbo-0125OpenAI |
1,226 1,219–1,233 |
| 306 | mistral-large-2402Mistral |
1,226 1,218–1,234 |
| 307 | Nova Micro 1.0Amazon |
1,223 1,214–1,231 |
| 308 | qwen1.5-110b-chatAlibaba |
1,219 1,209–1,228 |
| 308 | gemini-pro-dev-apiGoogle |
1,218 1,206–1,229 |
| 308 | hunyuan-standard-256kTencent |
1,208 1,188–1,229 |
| 310 | qwen1.5-72b-chatAlibaba |
1,214 1,206–1,223 |
| 310 | jamba-1.5-miniAI21 Labs |
1,214 1,202–1,227 |
| 310 | c4ai-aya-expanse-8bCohere |
1,213 1,202–1,224 |
| 310 | gemini-proGoogle |
1,208 1,189–1,227 |
| 311 | llama-3.1-tulu-3-8bAi2 |
1,200 1,181–1,220 |
| 314 | Mixtral 8x22B InstructMistral |
1,210 1,202–1,217 |
| 314 | command-rCohere |
1,209 1,201–1,217 |
| 314 | mistral-mediumMistral |
1,208 1,198–1,217 |
| 314 | reka-flash-21b-20240226-onlineRekaai |
1,204 1,192–1,216 |
| 316 | zephyr-orpo-141b-A35b-v0.1Huggingface |
1,196 1,178–1,214 |
| 318 | wizardlm-70bMicrosoft |
1,195 1,180–1,210 |
| 320 | reka-flash-21b-20240226Rekaai |
1,198 1,188–1,208 |
| 320 | gpt-3.5-turbo-1106OpenAI |
1,193 1,179–1,206 |
| 322 | falcon-180b-chatTII |
1,171 1,139–1,203 |
| 326 | gemma-2-2b-itGoogle |
1,193 1,187–1,200 |
| 326 | llama-3-8b-instructMeta |
1,192 1,185–1,199 |
| 326 | granite-3.1-8b-instructIBM |
1,180 1,159–1,200 |
| 327 | openchat-3.5Openchat |
1,177 1,161–1,192 |
| 328 | deepseek-llm-67b-chatDeepSeek |
1,171 1,152–1,189 |
| 329 | Llama 3.1 8B InstructMeta |
1,182 1,176–1,189 |
| 332 | dbrx-instruct-previewDatabricks |
1,176 1,167–1,186 |
| 333 | yi-1.5-34b-chat01 Ai |
1,175 1,166–1,184 |
| 333 | phi-3-medium-4k-instructMicrosoft |
1,173 1,164–1,182 |
| 333 | qwen1.5-32b-chatAlibaba |
1,172 1,162–1,183 |
| 333 | vicuna-33bLmsys |
1,171 1,161–1,182 |
| 333 | tulu-2-dpo-70bAi2 |
1,166 1,149–1,182 |
| 336 | mixtral-8x7b-instruct-v0.1Mistral |
1,171 1,163–1,178 |
| 337 | qwen1.5-14b-chatAlibaba |
1,165 1,153–1,176 |
| 337 | yi-34b-chat01 Ai |
1,164 1,153–1,176 |
| 337 | openchat-3.5-0106Openchat |
1,164 1,151–1,176 |
| 337 | openhermes-2.5-mistral-7bTeknium |
1,159 1,141–1,177 |
| 338 | snowflake-arctic-instructSnowflake |
1,160 1,150–1,170 |
| 338 | wizardlm-13bMicrosoft |
1,153 1,138–1,168 |
| 341 | internlm2_5-20b-chatShanghai Ai Lab |
1,152 1,139–1,164 |
| 342 | gemma-1.1-7b-itGoogle |
1,154 1,145–1,163 |
| 342 | mpt-30b-chatMosaicml |
1,140 1,117–1,162 |
| 345 | granite-3.0-8b-instructIBM |
1,144 1,129–1,160 |
| 345 | nous-hermes-2-mixtral-8x7b-dpoNousresearch |
1,141 1,121–1,160 |
| 345 | solar-10.7b-instruct-v1.0Upstage |
1,140 1,119–1,160 |
| 345 | dolphin-2.2.1-mistral-7bCognitivecomputations |
1,132 1,103–1,160 |
| 346 | vicuna-13bLmsys |
1,146 1,135–1,156 |
| 346 | Llama 3.2 3B InstructMeta |
1,143 1,128–1,157 |
| 347 | starling-lm-7b-alphaBerkeley |
1,140 1,127–1,153 |
| 349 | phi-3-small-8k-instructMicrosoft |
1,141 1,131–1,151 |
| 351 | llama2-70b-steerlm-chatNVIDIA |
1,128 1,106–1,150 |
| 352 | granite-3.1-2b-instructIBM |
1,128 1,107–1,149 |
| 352 | guanaco-33bTimdettmers |
1,123 1,100–1,146 |
| 352 | zephyr-7b-alphaHuggingface |
1,123 1,098–1,148 |
| 354 | qwen-14b-chatAlibaba |
1,121 1,103–1,140 |
| 355 | llama-2-70b-chatMeta |
1,131 1,122–1,139 |
| 355 | starling-lm-7b-betaNexusflow |
1,127 1,115–1,139 |
| 358 | zephyr-7b-betaHuggingface |
1,121 1,107–1,135 |
| 358 | vicuna-7bLmsys |
1,116 1,101–1,132 |
| 358 | qwq-32b-previewAlibaba |
1,113 1,092–1,134 |
| 358 | qwen1.5-7b-chatAlibaba |
1,113 1,095–1,131 |
| 362 | mistral-7b-instruct-v0.2Mistral |
1,113 1,103–1,124 |
| 362 | granite-3.0-2b-instructIBM |
1,106 1,091–1,122 |
| 364 | gemma-7b-itGoogle |
1,105 1,091–1,120 |
| 364 | codellama-34b-instructMeta |
1,105 1,090–1,119 |
| 366 | llama-2-13b-chatMeta |
1,104 1,094–1,115 |
| 366 | phi-3-mini-4k-instruct-june-2024Microsoft |
1,104 1,092–1,115 |
| 367 | phi-3-mini-128k-instructMicrosoft |
1,099 1,087–1,111 |
| 367 | palm-2Google |
1,097 1,082–1,112 |
| 367 | mistral-7b-instructMistral |
1,095 1,080–1,111 |
| 367 | stripedhyena-nous-7bTogether |
1,093 1,075–1,111 |
| 370 | phi-3-mini-4k-instructMicrosoft |
1,093 1,082–1,103 |
| 370 | gemma-1.1-2b-itGoogle |
1,092 1,079–1,105 |
| 370 | smollm2-1.7b-instructHuggingface |
1,079 1,053–1,106 |
| 375 | alpaca-13bStanford |
1,081 1,062–1,100 |
| 378 | gpt4all-13b-snoozyNomic |
1,062 1,031–1,093 |
| 380 | qwen1.5-4b-chatAlibaba |
1,077 1,062–1,092 |
| 380 | Llama 3.2 1B InstructMeta |
1,076 1,061–1,091 |
| 380 | codellama-70b-instructMeta |
1,055 1,018–1,092 |
| 383 | llama-2-7b-chatMeta |
1,076 1,064–1,087 |
| 388 | gemma-2b-itGoogle |
1,058 1,039–1,077 |
| 388 | mpt-7b-chatMosaicml |
1,057 1,037–1,078 |
| 389 | koala-13bBerkeley |
1,049 1,031–1,067 |
| 392 | chatglm3-6bZ.ai |
1,041 1,021–1,061 |
| 394 | olmo-7b-instructAi2 |
1,023 1,004–1,041 |
| 395 | chatglm2-6bZ.ai |
1,016 994–1,039 |
| 397 | oasst-pythia-12bOpenassistant |
1,013 995–1,031 |
| 398 | RWKV-4-Raven-14BRwkv |
1,003 984–1,023 |
| 399 | fastchat-t5-3bLmsys |
998 977–1,019 |
| 401 | chatglm-6bZ.ai |
982 961–1,002 |
| 401 | dolly-v2-12bDatabricks |
975 952–998 |
| 406 | stablelm-tuned-alpha-7bStabilityai |
931 907–955 |
| 406 | llama-13bMeta |
929 900–957 |
Models share a rank when their ranges overlap. Results as published by Arena (formerly LMArena); we do not re-run them.
What it measures
Human preference on writing and language prompts.
What it does not measure
Not accuracy or correctness: it ranks which answer voters preferred, with answer length and formatting controlled for.
Source
Contains data from the Arena Leaderboard Dataset by Arena, licensed under CC BY 4.0. Licence: Creative Commons Attribution 4.0 International.