1,519
1,496–1,542
Benchmarks / Arena (formerly LMArena)
Reported by Arena (formerly LMArena)
Arena (formerly LMArena)
Human preference on business, management and finance prompts.
- Results dated
- 25 Sep 2026
- Models
- 402
- Unit
- Arena rating
Full results
| # | Model | Arena Text: business, management and finance · 95% range Arena rating, higher is better |
|---|---|---|
| 1 | Muse Spark 1.2Meta | |
| 1 | Claude Opus 4.7 (high)Anthropic |
1,508 1,501–1,514 |
| 1 | Muse Spark 1.3 (max)Meta |
1,506 1,492–1,519 |
| 1 | Claude Fable 5 (high)Anthropic |
1,503 1,496–1,511 |
| 1 | Claude Opus 4.6 (high)Anthropic |
1,501 1,495–1,507 |
| 1 | Claude Opus 4.6Anthropic |
1,501 1,495–1,507 |
| 1 | Claude Opus 5.5 (high)Anthropic |
1,499 1,471–1,528 |
| 1 | Claude Opus 4.7Anthropic |
1,495 1,489–1,502 |
| 2 | Muse Spark 1.1Meta |
1,493 1,485–1,501 |
| 2 | Kimi K3 (max)Moonshot AI |
1,489 1,480–1,498 |
| 2 | muse-sparkMeta |
1,489 1,478–1,501 |
| 2 | Claude Fable 5.1 (max)Anthropic |
1,485 1,471–1,499 |
| 2 | GPT-6 Astra (max)OpenAI |
1,483 1,466–1,499 |
| 2 | DeepSeek V4.1 Flash (max)DeepSeek |
1,481 1,464–1,498 |
| 3 | Claude Opus 4.8 (high)Anthropic |
1,490 1,483–1,496 |
| 3 | GPT-5.5 (high)OpenAI |
1,489 1,483–1,496 |
| 6 | Claude Opus 4.8Anthropic |
1,487 1,481–1,493 |
| 6 | Claude Opus 5 (high)Anthropic |
1,486 1,479–1,493 |
| 6 | GPT-5.6 Sol (xhigh)OpenAI |
1,485 1,477–1,493 |
| 6 | gpt-5.2-chat-latest-20260210OpenAI |
1,485 1,477–1,493 |
| 6 | Gemini 3.7 Flash (high)Google |
1,483 1,473–1,494 |
| 7 | GPT-5.5OpenAI |
1,486 1,479–1,492 |
| 7 | Claude Opus 5 (max)Anthropic |
1,482 1,473–1,491 |
| 7 | Gemini 3.8 Flash (high)Google |
1,482 1,472–1,492 |
| 7 | Qwen3.8 Max (0902)Alibaba |
1,480 1,470–1,490 |
| 8 | Claude Sonnet 4.6Anthropic |
1,482 1,476–1,488 |
| 8 | GPT-5.4 (high)OpenAI |
1,482 1,475–1,488 |
| 8 | GLM 5.3 (max)Z.ai |
1,474 1,463–1,486 |
| 9 | gpt-5.5-instantOpenAI |
1,476 1,467–1,485 |
| 9 | ernie-5.1Baidu |
1,475 1,468–1,483 |
| 9 | GPT-6 Sol (max)OpenAI |
1,465 1,445–1,484 |
| 9 | MiMo-V2.6-ProXiaomi |
1,462 1,440–1,484 |
| 11 | Gemini 3.1 Pro PreviewGoogle |
1,477 1,472–1,482 |
| 11 | GPT-5.4OpenAI |
1,475 1,468–1,481 |
| 11 | gemini-3-proGoogle |
1,474 1,467–1,481 |
| 11 | DeepSeek V4 Pro 0813DeepSeek |
1,467 1,453–1,481 |
| 11 | MiMo-V2.6-FlashXiaomi |
1,460 1,439–1,481 |
| 13 | Gemini 3.6 Flash (high)Google |
1,472 1,464–1,480 |
| 14 | qwen3.7-max-previewAlibaba |
1,459 1,439–1,479 |
| 15 | Claude Opus 4.5Anthropic |
1,472 1,466–1,478 |
| 15 | grok-4.20-beta1xAI |
1,470 1,461–1,478 |
| 16 | MiMo-V2.5-ProXiaomi |
1,471 1,465–1,478 |
| 18 | gemini-3-flashGoogle |
1,469 1,461–1,477 |
| 18 | muse-glimmerMeta |
1,455 1,433–1,477 |
| 20 | Gemini 3.5 Flash (high)Google |
1,468 1,461–1,475 |
| 20 | Claude Opus 4.5Anthropic |
1,467 1,460–1,475 |
| 20 | GPT-5.6 Terra (xhigh)OpenAI |
1,466 1,458–1,474 |
| 20 | GLM 5.3 FlashZ.ai |
1,463 1,453–1,474 |
| 22 | grok-4.20-beta-0309-reasoningxAI |
1,466 1,460–1,472 |
| 22 | Grok 4.5xAI |
1,465 1,457–1,473 |
| 22 | Claude Sonnet 5 (high)Anthropic |
1,465 1,458–1,472 |
| 22 | GPT-6 Luna (max)OpenAI |
1,452 1,432–1,473 |
| 24 | GLM 5.1Z.ai |
1,465 1,459–1,472 |
| 24 | GPT-5.6 Luna (xhigh)OpenAI |
1,463 1,455–1,471 |
| 25 | qwen3.5-max-previewAlibaba |
1,462 1,453–1,471 |
| 26 | GLM 5.2 (max)Z.ai |
1,463 1,456–1,471 |
| 27 | dola-seed-2.0-proBytedance |
1,464 1,458–1,470 |
| 27 | Gemini 3.5 Flash (medium)Google |
1,461 1,454–1,469 |
| 27 | Qwen3.6 Max PreviewAlibaba |
1,450 1,432–1,469 |
| 29 | gpt-5.3-chat-latestOpenAI |
1,459 1,452–1,467 |
| 30 | grok-4.1xAI |
1,461 1,455–1,467 |
| 31 | DeepSeek V4 Pro 0423DeepSeek |
1,460 1,453–1,466 |
| 33 | Claude Sonnet 4.5Anthropic |
1,461 1,455–1,466 |
| 33 | GPT-5.4 Mini (high)OpenAI |
1,460 1,453–1,466 |
| 33 | Gemma 4 31BGoogle |
1,449 1,433–1,466 |
| 34 | grok-4.1-thinkingxAI |
1,459 1,453–1,465 |
| 35 | Claude Sonnet 4.5Anthropic |
1,459 1,453–1,464 |
| 37 | Kimi K2.6Moonshot AI |
1,455 1,447–1,462 |
| 37 | Gemini 3.5 Flash LiteGoogle |
1,454 1,445–1,462 |
| 37 | mimo-v2-proXiaomi |
1,454 1,445–1,462 |
| 37 | Gemma 4 26B A4BGoogle |
1,444 1,427–1,461 |
| 38 | Hy3Tencent |
1,447 1,433–1,461 |
| 39 | grok-4.20-multi-agent-beta-0309xAI |
1,454 1,448–1,461 |
| 40 | gemini-3-flash (minimal reasoning)Google |
1,455 1,449–1,460 |
| 42 | GLM 5Z.ai |
1,451 1,443–1,459 |
| 43 | chatgpt-4o-latest-20250326OpenAI |
1,453 1,447–1,458 |
| 44 | amazon-nova-experimental-chat-26-02-10Amazon |
1,435 1,413–1,458 |
| 47 | MiniMax M3MiniMax |
1,450 1,443–1,456 |
| 47 | Grok 4.6 (high)xAI |
1,447 1,438–1,457 |
| 48 | GPT-5.1 (high)OpenAI |
1,448 1,441–1,456 |
| 48 | ernie-5.0-preview-1203Baidu |
1,442 1,429–1,456 |
| 51 | deepseek-v4-pro-preview (high reasoning)DeepSeek |
1,448 1,441–1,455 |
| 51 | GLM 4.7Z.ai |
1,442 1,430–1,455 |
| 52 | Claude Opus 4.1Anthropic |
1,447 1,441–1,454 |
| 52 | qwen3-max-previewAlibaba |
1,445 1,436–1,454 |
| 52 | Qwen3 VL 235B A22B InstructAlibaba |
1,440 1,426–1,454 |
| 59 | Claude Opus 4.1Anthropic |
1,447 1,441–1,452 |
| 59 | Qwen3.6 PlusAlibaba |
1,446 1,439–1,453 |
| 59 | Qwen3.7 PlusAlibaba |
1,445 1,437–1,452 |
| 59 | InklingThinkingmachines |
1,444 1,436–1,452 |
| 59 | GLM 5V TurboZ.ai |
1,440 1,427–1,453 |
| 60 | Qwen3.5 397B A17BAlibaba |
1,445 1,440–1,451 |
| 60 | GPT-5.2OpenAI |
1,444 1,438–1,450 |
| 60 | GPT-5.2 (high)OpenAI |
1,443 1,437–1,450 |
| 60 | GPT-5.1OpenAI |
1,443 1,436–1,450 |
| 60 | mimo-v2-omniXiaomi |
1,441 1,431–1,451 |
| 61 | gpt-5-chatOpenAI |
1,440 1,432–1,449 |
| 62 | MiMo-V2.5Xiaomi |
1,441 1,433–1,448 |
| 62 | ernie-5.0-0110Baidu |
1,441 1,433–1,448 |
| 64 | Kimi K2.5Moonshot AI |
1,441 1,435–1,446 |
| 64 | Grok 4.3xAI |
1,440 1,433–1,446 |
| 64 | DeepSeek V4 Flash 0423DeepSeek |
1,439 1,432–1,445 |
| 64 | DeepSeek V3.1 TerminusDeepSeek |
1,423 1,401–1,445 |
| 65 | Qwen3.8 27BAlibaba |
1,434 1,423–1,445 |
| 65 | Grok 4.7 (xhigh)xAI |
1,424 1,403–1,445 |
| 67 | Mistral Medium 3.5Mistral |
1,430 1,417–1,443 |
| 69 | Gemini 2.5 ProGoogle |
1,437 1,433–1,442 |
| 69 | amazon-nova-experimental-chat-26-01-10Amazon |
1,420 1,398–1,442 |
| 69 | hunyuan-vision-1.5-thinkingTencent |
1,415 1,388–1,442 |
| 71 | deepseek-v4-flash-preview (high reasoning)DeepSeek |
1,434 1,427–1,441 |
| 71 | longcat-flash-chat-2602-expMeituan |
1,433 1,424–1,441 |
| 71 | gpt-4.5-preview-2025-02-27OpenAI |
1,426 1,410–1,441 |
| 73 | Kimi K2.5Moonshot AI |
1,426 1,411–1,441 |
| 81 | Qwen3 235B A22B Instruct 2507Alibaba |
1,432 1,427–1,437 |
| 86 | DeepSeek V3.2 ExpDeepSeek |
1,422 1,409–1,434 |
| 89 | o3OpenAI |
1,426 1,420–1,433 |
| 91 | Qwen3 MaxAlibaba |
1,419 1,405–1,433 |
| 98 | nvidia-nemotron-3-ultra-550b-a55b-nvfp4NVIDIA |
1,418 1,405–1,431 |
| 98 | DeepSeek V3.1DeepSeek |
1,417 1,404–1,431 |
| 98 | longcat-flash-chatMeituan |
1,417 1,405–1,430 |
| 98 | ernie-5.0-preview-1022Baidu |
1,412 1,394–1,431 |
| 99 | MiniMax M2.7MiniMax |
1,423 1,417–1,429 |
| 99 | DeepSeek V3.2DeepSeek |
1,423 1,416–1,430 |
| 99 | Qwen3.5-122B-A10BAlibaba |
1,422 1,413–1,430 |
| 100 | DeepSeek V3.2DeepSeek |
1,422 1,415–1,428 |
| 103 | Hy3 previewTencent |
1,411 1,395–1,427 |
| 104 | kimi-k2-thinking-turboMoonshot AI |
1,421 1,415–1,427 |
| 104 | Gemini 3.1 Flash Lite PreviewGoogle |
1,421 1,414–1,427 |
| 105 | DeepSeek V3.1DeepSeek |
1,414 1,402–1,426 |
| 105 | amazon-nova-experimental-chat-12-10Amazon |
1,402 1,379–1,425 |
| 106 | Claude Haiku 4.5Anthropic |
1,420 1,415–1,424 |
| 106 | grok-4-1-fast-reasoningxAI |
1,417 1,411–1,423 |
| 107 | GPT-4.1OpenAI |
1,415 1,408–1,422 |
| 107 | Qwen3 235B A22BAlibaba |
1,414 1,406–1,423 |
| 107 | GPT-5 (high)OpenAI |
1,413 1,405–1,422 |
| 107 | GLM 4.5Z.ai |
1,412 1,403–1,421 |
| 107 | Qwen3 Next 80B A3B InstructAlibaba |
1,411 1,402–1,421 |
| 107 | DeepSeek V3.2 ExpDeepSeek |
1,406 1,392–1,420 |
| 107 | hunyuan-turbos-20250416Tencent |
1,406 1,390–1,422 |
| 108 | GLM 4.6Z.ai |
1,412 1,404–1,420 |
| 108 | claude-opus-4 (thinking-16k reasoning)Anthropic |
1,409 1,401–1,418 |
| 108 | Kimi K2 0711Moonshot AI |
1,409 1,400–1,418 |
| 108 | R1 0528DeepSeek |
1,407 1,395–1,418 |
| 108 | Qwen3 235B A22B Thinking 2507Alibaba |
1,404 1,389–1,419 |
| 108 | grok-4-fast-chatxAI |
1,401 1,383–1,418 |
| 108 | DeepSeek V3.1 TerminusDeepSeek |
1,397 1,375–1,418 |
| 110 | Qwen3.5-27BAlibaba |
1,408 1,400–1,416 |
| 110 | Inkling SmallThinkingmachines |
1,408 1,398–1,417 |
| 110 | Kimi K2 0905Moonshot AI |
1,404 1,391–1,417 |
| 110 | Gemma 3 12BGoogle |
1,386 1,356–1,416 |
| 112 | Mistral Large 3 2512Mistral |
1,409 1,404–1,415 |
| 117 | Mistral Medium 3.1Mistral |
1,407 1,402–1,412 |
| 117 | GPT-5.4 Nano (high)OpenAI |
1,405 1,398–1,412 |
| 117 | claude-opus-4Anthropic |
1,403 1,396–1,411 |
| 117 | gemini-2.5-flash-preview-09-2025Google |
1,403 1,395–1,411 |
| 117 | grok-3-preview-02-24xAI |
1,403 1,393–1,413 |
| 117 | grok-4-fast-reasoningxAI |
1,403 1,393–1,413 |
| 119 | Qwen3.5-35B-A3BAlibaba |
1,403 1,395–1,411 |
| 121 | Qwen3.5-FlashAlibaba |
1,403 1,397–1,409 |
| 122 | MiniMax M2.5MiniMax |
1,400 1,393–1,407 |
| 122 | Qwen3 30B A3B Instruct 2507Alibaba |
1,397 1,388–1,407 |
| 122 | Qwen3 VL 235B A22B ThinkingAlibaba |
1,392 1,378–1,407 |
| 130 | Gemini 2.5 FlashGoogle |
1,399 1,395–1,404 |
| 132 | grok-4-0709xAI |
1,395 1,388–1,403 |
| 134 | R1DeepSeek |
1,388 1,374–1,402 |
| 137 | Qwen3 Coder 480B A35BAlibaba |
1,391 1,382–1,401 |
| 137 | mimo-v2-flash (thinking reasoning)Xiaomi |
1,388 1,376–1,401 |
| 137 | hunyuan-turbos-20250226Tencent |
1,368 1,337–1,400 |
| 139 | mimo-v2-flash (no reasoning)Xiaomi |
1,393 1,386–1,400 |
| 142 | Step 3.5 FlashStepfun |
1,390 1,384–1,397 |
| 142 | DeepSeek V3 0324DeepSeek |
1,390 1,382–1,397 |
| 142 | Solar Pro 4Upstage |
1,386 1,374–1,398 |
| 143 | Mistral Medium 3Mistral |
1,387 1,378–1,396 |
| 147 | Claude Sonnet 4Anthropic |
1,387 1,379–1,395 |
| 147 | GLM 4.6VZ.ai |
1,370 1,345–1,395 |
| 150 | amazon-nova-experimental-chat-10-09Amazon |
1,370 1,346–1,394 |
| 153 | minimax-m2.1-previewMiniMax |
1,382 1,371–1,393 |
| 156 | trinity-large-previewArcee Ai |
1,381 1,373–1,389 |
| 156 | Claude Sonnet 4Anthropic |
1,381 1,372–1,389 |
| 156 | GPT-5 Mini (high)OpenAI |
1,381 1,372–1,390 |
| 160 | GPT-4.1 MiniOpenAI |
1,379 1,371–1,387 |
| 160 | o4 MiniOpenAI |
1,379 1,372–1,387 |
| 160 | hunyuan-t1-20250711Tencent |
1,366 1,345–1,388 |
| 161 | gemini-2.5-flash-lite-preview-09-2025 (no reasoning)Google |
1,378 1,371–1,385 |
| 161 | Gemma 3 27BGoogle |
1,377 1,369–1,385 |
| 161 | amazon-nova-experimental-chat-11-10Amazon |
1,377 1,368–1,385 |
| 164 | o1OpenAI |
1,370 1,358–1,382 |
| 164 | intellect-3Primeintellect |
1,363 1,345–1,382 |
| 164 | hunyuan-turbo-0110Tencent |
1,348 1,315–1,382 |
| 165 | GLM 4.5 AirZ.ai |
1,372 1,364–1,380 |
| 165 | Qwen3 235B A22BAlibaba |
1,370 1,361–1,380 |
| 165 | GLM 4.7 FlashZ.ai |
1,368 1,356–1,380 |
| 167 | qwen2.5-maxAlibaba |
1,369 1,359–1,378 |
| 168 | Qwen3 Next 80B A3B ThinkingAlibaba |
1,367 1,355–1,378 |
| 169 | claude-3-7-sonnet-20250219Anthropic |
1,368 1,359–1,376 |
| 169 | llama-3.1-nemotron-ultra-253b-v1NVIDIA |
1,342 1,308–1,377 |
| 170 | Trinity Large ThinkingArcee Ai |
1,367 1,359–1,375 |
| 170 | amazon-nova-experimental-chat-10-20Amazon |
1,363 1,350–1,376 |
| 171 | step-3Stepfun |
1,358 1,341–1,375 |
| 173 | Mistral Small 3.2 24BMistral |
1,363 1,352–1,374 |
| 175 | claude-3-7-sonnet-20250219 (thinking-32k reasoning)Anthropic |
1,363 1,355–1,372 |
| 178 | gemini-2.5-flash-lite-preview-06-17 (thinking reasoning)Google |
1,363 1,354–1,371 |
| 180 | gemini-2.0-flash-lite-preview-02-05Google |
1,359 1,347–1,370 |
| 181 | glm-4-plus-0111Z.ai |
1,345 1,321–1,369 |
| 182 | nvidia-nemotron-3.5-lightning-30b-a3b-nvfp4NVIDIA |
1,352 1,339–1,365 |
| 182 | Nova 2 LiteAmazon |
1,351 1,339–1,364 |
| 182 | Nemotron 3 SuperNVIDIA |
1,350 1,335–1,365 |
| 182 | MiniMax M2MiniMax |
1,348 1,331–1,364 |
| 182 | Qwen-PlusAlibaba |
1,342 1,318–1,365 |
| 182 | Qwen3 32BAlibaba |
1,340 1,314–1,366 |
| 182 | Mercury 2Inception |
1,340 1,316–1,364 |
| 183 | o1-previewOpenAI |
1,353 1,343–1,364 |
| 183 | olmo-3.1-32b-instructAi2 |
1,350 1,337–1,363 |
| 183 | DeepSeek V3DeepSeek |
1,350 1,338–1,362 |
| 183 | o3 Mini HighOpenAI |
1,349 1,336–1,363 |
| 183 | ling-flash-2.0Ant Group |
1,347 1,331–1,363 |
| 183 | GLM 4.5VZ.ai |
1,342 1,323–1,362 |
| 184 | claude-3-5-sonnet-20241022Anthropic |
1,354 1,348–1,361 |
| 184 | gpt-oss-120bOpenAI |
1,351 1,343–1,360 |
| 184 | MiniMax M1MiniMax |
1,351 1,343–1,359 |
| 184 | Gemma 3 4BGoogle |
1,330 1,300–1,360 |
| 185 | grok-3-mini-betaxAI |
1,349 1,338–1,359 |
| 188 | Command ACohere |
1,351 1,345–1,358 |
| 189 | gemini-2.0-flash-001Google |
1,348 1,339–1,356 |
| 189 | grok-3-mini (high reasoning)xAI |
1,344 1,333–1,356 |
| 190 | granite-4.2-30bIBM |
1,331 1,307–1,356 |
| 193 | GPT-5 Nano (high)OpenAI |
1,339 1,323–1,354 |
| 193 | nvidia-llama-3.3-nemotron-super-49b-v1.5NVIDIA |
1,329 1,306–1,352 |
| 193 | mercuryInception |
1,321 1,289–1,353 |
| 193 | llama-3.3-nemotron-49b-super-v1NVIDIA |
1,319 1,285–1,354 |
| 196 | qwq-32bAlibaba |
1,337 1,327–1,347 |
| 202 | GPT-4.1 NanoOpenAI |
1,321 1,299–1,343 |
| 205 | ring-flash-2.0Ant Group |
1,326 1,309–1,342 |
| 206 | gemini-1.5-pro-002Google |
1,331 1,323–1,339 |
| 206 | Qwen3 30B A3BAlibaba |
1,330 1,320–1,339 |
| 206 | step-1o-turbo-202506Stepfun |
1,324 1,309–1,340 |
| 212 | gpt-oss-20bOpenAI |
1,323 1,309–1,337 |
| 213 | step-2-16k-exp-202412Stepfun |
1,312 1,287–1,336 |
| 214 | o3 MiniOpenAI |
1,328 1,321–1,335 |
| 216 | qwen2.5-plus-1127Alibaba |
1,315 1,298–1,332 |
| 218 | GPT-4o (2024-05-13)OpenAI |
1,322 1,314–1,329 |
| 218 | claude-3-5-sonnet-20240620Anthropic |
1,321 1,313–1,329 |
| 218 | glm-4-plusZ.ai |
1,319 1,308–1,330 |
| 218 | yi-lightning01 Ai |
1,318 1,307–1,330 |
| 218 | gemma-3n-e4b-itGoogle |
1,317 1,306–1,328 |
| 218 | Nemotron 3 Nano 30B A3BNVIDIA |
1,317 1,306–1,328 |
| 218 | hunyuan-standard-2025-02-10Tencent |
1,300 1,272–1,329 |
| 219 | o1-miniOpenAI |
1,319 1,310–1,327 |
| 219 | Llama 4 MaverickMeta |
1,317 1,309–1,325 |
| 219 | Llama 4 ScoutMeta |
1,316 1,307–1,325 |
| 219 | gemini-advanced-0514Google |
1,314 1,303–1,324 |
| 219 | olmo-3-32b-thinkAi2 |
1,308 1,290–1,326 |
| 219 | deepseek-v2.5-1210DeepSeek |
1,306 1,284–1,327 |
| 222 | claude-3-5-haiku-20241022Anthropic |
1,315 1,308–1,322 |
| 222 | grok-2-2024-08-13xAI |
1,315 1,307–1,323 |
| 222 | llama-3.1-405b-instruct-fp8Meta |
1,314 1,306–1,323 |
| 222 | llama-3.1-405b-instruct-bf16Meta |
1,313 1,304–1,322 |
| 223 | athene-v2-chatNexusflow |
1,310 1,298–1,321 |
| 224 | hunyuan-large-2025-02-10Tencent |
1,290 1,259–1,320 |
| 225 | gemini-1.5-pro-001Google |
1,311 1,302–1,320 |
| 225 | GPT-4o-mini (2024-07-18)OpenAI |
1,311 1,303–1,319 |
| 225 | qwen-max-0919Alibaba |
1,306 1,293–1,319 |
| 225 | llama-3.1-nemotron-51b-instructNVIDIA |
1,291 1,264–1,319 |
| 226 | granite-4.1-8bIBM |
1,293 1,269–1,317 |
| 228 | llama-3.1-nemotron-70b-instructNVIDIA |
1,295 1,276–1,314 |
| 228 | Granite 4.2 8BIBM |
1,290 1,265–1,315 |
| 231 | claude-3-opus-20240229Anthropic |
1,305 1,298–1,312 |
| 231 | Llama 3.3 70B InstructMeta |
1,304 1,296–1,311 |
| 231 | GPT-4o (2024-08-06)OpenAI |
1,302 1,293–1,311 |
| 231 | deepseek-v2.5DeepSeek |
1,300 1,289–1,312 |
| 231 | athene-70b-0725Nexusflow |
1,298 1,285–1,312 |
| 232 | grok-2-mini-2024-08-13xAI |
1,301 1,293–1,309 |
| 232 | granite-4.2-3bIBM |
1,284 1,258–1,310 |
| 236 | Qwen2.5 72B InstructAlibaba |
1,299 1,290–1,309 |
| 237 | gemini-1.5-flash-002Google |
1,298 1,289–1,308 |
| 237 | llama-3.1-tulu-3-70bAi2 |
1,276 1,244–1,308 |
| 247 | Mistral Small 3.1 24BMistral |
1,297 1,288–1,306 |
| 247 | mistral-large-2411Mistral |
1,294 1,283–1,305 |
| 247 | magistral-medium-2506Mistral |
1,292 1,278–1,305 |
| 247 | Qwen2.5 Coder 32B InstructAlibaba |
1,283 1,261–1,305 |
| 251 | GPT-4 TurboOpenAI |
1,293 1,284–1,301 |
| 251 | gemma-2-9b-it-simpoPrinceton Nlp |
1,284 1,267–1,302 |
| 251 | reka-core-20240904Rekaai |
1,280 1,261–1,300 |
| 252 | Mistral Large 2407Mistral |
1,290 1,281–1,300 |
| 252 | hunyuan-large-visionTencent |
1,278 1,257–1,299 |
| 256 | olmo-3.1-32b-thinkAi2 |
1,280 1,264–1,296 |
| 257 | Llama 3.1 70B InstructMeta |
1,287 1,279–1,296 |
| 257 | gemini-1.5-flash-001Google |
1,284 1,275–1,293 |
| 260 | gpt-4-0125-previewOpenAI |
1,283 1,274–1,292 |
| 260 | Nova Pro 1.0Amazon |
1,279 1,268–1,290 |
| 260 | ibm-granite-h-smallIBM |
1,272 1,252–1,291 |
| 266 | gpt-4-1106-previewOpenAI |
1,279 1,271–1,288 |
| 266 | Command R+ (08-2024)Cohere |
1,271 1,254–1,288 |
| 267 | Gemma 2 27BGoogle |
1,278 1,271–1,286 |
| 268 | reka-flash-20240904Rekaai |
1,266 1,247–1,285 |
| 272 | Nova Lite 1.0Amazon |
1,269 1,257–1,282 |
| 272 | jamba-1.5-largeAI21 Labs |
1,264 1,246–1,283 |
| 274 | c4ai-aya-expanse-32bCohere |
1,268 1,257–1,279 |
| 276 | Mistral Small 3Mistral |
1,259 1,244–1,275 |
| 276 | llama-3.1-tulu-3-8bAi2 |
1,244 1,212–1,276 |
| 279 | nemotron-4-340b-instructNVIDIA |
1,258 1,245–1,271 |
| 281 | claude-3-sonnet-20240229Anthropic |
1,261 1,252–1,270 |
| 281 | gemma-2-9b-itGoogle |
1,261 1,252–1,269 |
| 282 | command-r-plusCohere |
1,259 1,250–1,269 |
| 282 | Phi 4Microsoft |
1,256 1,244–1,268 |
| 283 | olmo-2-0325-32b-instructAi2 |
1,236 1,204–1,267 |
| 285 | gemini-1.5-flash-8b-001Google |
1,255 1,245–1,265 |
| 285 | deepseek-coder-v2DeepSeek |
1,250 1,236–1,264 |
| 287 | glm-4-0520Z.ai |
1,245 1,228–1,262 |
| 290 | ministral-8b-2410Mistral |
1,235 1,212–1,259 |
| 290 | granite-3.1-8b-instructIBM |
1,225 1,193–1,258 |
| 291 | llama-3-70b-instructMeta |
1,249 1,241–1,257 |
| 294 | claude-3-haiku-20240307Anthropic |
1,248 1,240–1,256 |
| 294 | gpt-4-0314OpenAI |
1,245 1,234–1,256 |
| 294 | hunyuan-standard-256kTencent |
1,224 1,193–1,255 |
| 300 | Nova Micro 1.0Amazon |
1,233 1,220–1,247 |
| 300 | Command R (08-2024)Cohere |
1,229 1,211–1,246 |
| 301 | c4ai-aya-expanse-8bCohere |
1,228 1,210–1,246 |
| 304 | gemini-pro-dev-apiGoogle |
1,227 1,210–1,244 |
| 306 | jamba-1.5-miniAI21 Labs |
1,223 1,205–1,242 |
| 306 | granite-3.1-2b-instructIBM |
1,213 1,184–1,243 |
| 307 | gpt-4-0613OpenAI |
1,231 1,222–1,241 |
| 307 | reka-flash-21b-20240226-onlineRekaai |
1,224 1,209–1,240 |
| 308 | qwen2-72b-instructAlibaba |
1,227 1,217–1,238 |
| 310 | command-rCohere |
1,223 1,213–1,234 |
| 310 | mistral-large-2402Mistral |
1,220 1,210–1,230 |
| 310 | qwen1.5-110b-chatAlibaba |
1,219 1,207–1,232 |
| 311 | reka-flash-21b-20240226Rekaai |
1,213 1,200–1,226 |
| 312 | Mixtral 8x22B InstructMistral |
1,211 1,201–1,221 |
| 313 | Llama 3.1 8B InstructMeta |
1,210 1,201–1,219 |
| 313 | mistral-mediumMistral |
1,206 1,194–1,219 |
| 313 | qwen1.5-72b-chatAlibaba |
1,206 1,194–1,217 |
| 314 | llama-3-8b-instructMeta |
1,205 1,196–1,214 |
| 318 | zephyr-orpo-141b-A35b-v0.1Huggingface |
1,183 1,155–1,210 |
| 322 | gpt-3.5-turbo-0125OpenAI |
1,198 1,189–1,208 |
| 322 | yi-1.5-34b-chat01 Ai |
1,195 1,183–1,208 |
| 322 | llama2-70b-steerlm-chatNVIDIA |
1,172 1,137–1,207 |
| 323 | gemini-proGoogle |
1,177 1,147–1,207 |
| 329 | qwen1.5-32b-chatAlibaba |
1,182 1,169–1,196 |
| 329 | wizardlm-70bMicrosoft |
1,171 1,147–1,195 |
| 333 | gemma-2-2b-itGoogle |
1,181 1,172–1,190 |
| 333 | phi-3-medium-4k-instructMicrosoft |
1,179 1,166–1,191 |
| 333 | openchat-3.5-0106Openchat |
1,171 1,153–1,189 |
| 333 | tulu-2-dpo-70bAi2 |
1,165 1,140–1,191 |
| 334 | mixtral-8x7b-instruct-v0.1Mistral |
1,177 1,168–1,187 |
| 334 | dbrx-instruct-previewDatabricks |
1,174 1,161–1,187 |
| 334 | qwen1.5-14b-chatAlibaba |
1,170 1,154–1,186 |
| 334 | yi-34b-chat01 Ai |
1,169 1,153–1,186 |
| 334 | starling-lm-7b-betaNexusflow |
1,169 1,153–1,186 |
| 334 | deepseek-llm-67b-chatDeepSeek |
1,160 1,132–1,188 |
| 334 | nous-hermes-2-mixtral-8x7b-dpoNousresearch |
1,151 1,117–1,186 |
| 335 | gemma-1.1-7b-itGoogle |
1,170 1,157–1,183 |
| 335 | internlm2_5-20b-chatShanghai Ai Lab |
1,167 1,150–1,184 |
| 336 | llama-2-70b-chatMeta |
1,160 1,148–1,172 |
| 336 | phi-3-small-8k-instructMicrosoft |
1,159 1,145–1,173 |
| 336 | starling-lm-7b-alphaBerkeley |
1,158 1,138–1,179 |
| 336 | granite-3.0-8b-instructIBM |
1,150 1,127–1,172 |
| 336 | qwen1.5-7b-chatAlibaba |
1,148 1,120–1,176 |
| 337 | gpt-3.5-turbo-1106OpenAI |
1,150 1,131–1,169 |
| 337 | openchat-3.5Openchat |
1,146 1,122–1,169 |
| 340 | vicuna-33bLmsys |
1,150 1,135–1,166 |
| 340 | Llama 3.2 3B InstructMeta |
1,146 1,126–1,166 |
| 341 | granite-3.0-2b-instructIBM |
1,136 1,115–1,158 |
| 341 | openhermes-2.5-mistral-7bTeknium |
1,131 1,101–1,160 |
| 341 | solar-10.7b-instruct-v1.0Upstage |
1,125 1,091–1,159 |
| 341 | mpt-30b-chatMosaicml |
1,121 1,083–1,158 |
| 347 | snowflake-arctic-instructSnowflake |
1,139 1,126–1,153 |
| 347 | llama-2-13b-chatMeta |
1,134 1,118–1,150 |
| 347 | codellama-34b-instructMeta |
1,128 1,103–1,152 |
| 351 | mistral-7b-instruct-v0.2Mistral |
1,130 1,114–1,145 |
| 351 | qwen-14b-chatAlibaba |
1,118 1,090–1,146 |
| 352 | gemma-7b-itGoogle |
1,123 1,101–1,145 |
| 353 | qwq-32b-previewAlibaba |
1,108 1,076–1,140 |
| 356 | vicuna-13bLmsys |
1,118 1,101–1,135 |
| 358 | wizardlm-13bMicrosoft |
1,107 1,082–1,131 |
| 358 | gemma-2b-itGoogle |
1,101 1,073–1,129 |
| 361 | phi-3-mini-4k-instructMicrosoft |
1,112 1,098–1,125 |
| 361 | llama-2-7b-chatMeta |
1,105 1,087–1,122 |
| 362 | phi-3-mini-4k-instruct-june-2024Microsoft |
1,104 1,087–1,121 |
| 363 | gemma-1.1-2b-itGoogle |
1,099 1,080–1,118 |
| 364 | guanaco-33bTimdettmers |
1,080 1,042–1,118 |
| 365 | palm-2Google |
1,092 1,067–1,116 |
| 367 | phi-3-mini-128k-instructMicrosoft |
1,098 1,082–1,114 |
| 367 | zephyr-7b-betaHuggingface |
1,094 1,073–1,114 |
| 367 | olmo-7b-instructAi2 |
1,086 1,060–1,113 |
| 367 | qwen1.5-4b-chatAlibaba |
1,084 1,060–1,108 |
| 367 | vicuna-7bLmsys |
1,084 1,057–1,111 |
| 367 | stripedhyena-nous-7bTogether |
1,081 1,051–1,110 |
| 367 | smollm2-1.7b-instructHuggingface |
1,065 1,022–1,108 |
| 372 | Llama 3.2 1B InstructMeta |
1,070 1,050–1,091 |
| 372 | mistral-7b-instructMistral |
1,070 1,046–1,094 |
| 380 | koala-13bBerkeley |
1,048 1,019–1,077 |
| 381 | chatglm3-6bZ.ai |
1,044 1,012–1,075 |
| 390 | RWKV-4-Raven-14BRwkv |
1,009 975–1,042 |
| 390 | mpt-7b-chatMosaicml |
1,006 969–1,042 |
| 390 | chatglm2-6bZ.ai |
1,005 964–1,046 |
| 391 | fastchat-t5-3bLmsys |
1,000 962–1,038 |
| 393 | alpaca-13bStanford |
980 946–1,013 |
| 394 | oasst-pythia-12bOpenassistant |
960 930–991 |
| 394 | chatglm-6bZ.ai |
951 914–989 |
| 398 | stablelm-tuned-alpha-7bStabilityai |
906 866–946 |
| 399 | dolly-v2-12bDatabricks |
897 854–940 |
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 business, management and finance 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.