跳过主要内容
评测来源/Arena Agent Overall
观察中专业工作

Arena Agent Overall

LMArena · Arena Agent Overall 的公开模型评测结果。

官方评测
来源说明

该来源测量模型在「专业工作」领域的具体交付表现,原始成绩会和统一能力读数分开展示。

限制: 不同版本、推理设置和测试集可能导致分数变化,请结合数据时间阅读。

使用方式

参与排名的来源会影响能力分;参考、交叉参考和观察来源仅用于对比与核验。

本来源的模型成绩

Claude Opus 5.50.2014484011975185690.32026-11-04
Claude Sonnet 5.50.1542367475951577680.82026-11-04
GPT-6 Astra0.146970259628448179.42026-11-04
Claude Fable 5.10.1385700502851826877.72026-11-04
GPT-6.1 Sol0.1251394576790528275.02026-11-04
GPT-6.1 Terra0.1139604811320861972.82026-11-04
Gemini 4 Argon0.1052516482427223971.12026-11-04
Muse Wonder Full0.0942263674440960868.82026-11-04
Claude Fable 50.07822496321097565.62026-11-04
Claude Opus 50.0766195302361066765.32026-11-04
Claude Opus 4.80.0713137967656022964.32026-10-08
GPT-6 Sol0.0677813189292463563.62026-11-04
Muse Wonder0.04935621874618502459.92026-10-16
Muse Wonder0.04935621874618502459.92026-10-16
GPT-5.6 Sol0.0457226772956461559.12026-11-04
GPT-5.50.0312031395589905256.22026-10-20
MiMo V2.6 Pro0.02681036425687692655.42026-11-04
Grok 4.70.0268019230058959355.42026-11-04
MiMo V2.6 Flash0.0259833341973936355.22026-10-16
MiMo V2.6 Flash0.0259833341973936355.22026-10-16
Gemini 3.8 Flash0.02581951465300599555.22026-11-04
Muse Spark 1.30.02511383272877219855.02026-11-04
Claude Sonnet 50.02442968007107667554.92026-11-04
Muse Wonder0.02433580988964449754.92026-11-04
DeepSeek V4.1 Flash0.01912171809265232353.82026-10-28
GPT-6 Luna0.01893723903909208253.82026-11-04
Claude Haiku 5.50.01494437423166636853.02026-10-16
Claude Haiku 5.50.01494437423166636853.02026-10-16
GPT-5.40.0135975787735884652.72026-10-20
Grok 4.50.01194206456582642852.42026-10-08
GLM 5.30.01089092033800851552.22026-11-04
Kimi K30.00841945388519349951.72026-11-04
GLM-5.20.006386395044794314551.32026-11-04
Claude Haiku 5.50.00624418889898906551.22026-11-04
Qwen3.8 Max0.00478829314273428351.02026-11-04
Grok 4.60.00454262611550124750.92026-11-04
MiMo V2.6 Flash0.00200058939545213550.42026-11-04
GLM-5.3 Flash-0.01162085398877742847.72026-10-16
GLM-5.3 Flash-0.01162085398877742847.72026-10-16
GPT-5.6 Terra-0.0150469083457740847.02026-11-04
GPT-5.6 Luna-0.02332254736227054245.32026-10-20
GLM-5.3 Flash-0.02393965049615807645.22026-11-04
MiniMax M3.1 Flash-0.0284351299109260544.32026-11-04
Step 5 Preview-0.02876123878830287244.22026-11-04
Gemini 3.1 Pro Preview-0.04440981849728723641.12026-11-04
Mistral Large 4-0.0546265523720927239.12026-11-04
Qwen3.7 Max-0.065199604617491537.02026-11-04
Qwen3.7 Plus-0.0745994608608022335.12026-11-04
Muse Spark 1.1-0.0765945953643827534.72026-11-04
Muse Spark 1.2-0.0769898845975877534.62026-11-04
Hy3-0.077196984266655834.62026-11-04
Gemini 3.6 Flash-0.077918709620892334.42026-10-08
MiMo-V2.5-Pro-0.0902703287965474831.92026-11-04
MiniMax-M3-0.1039670705573458629.22026-11-04
NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4-0.159321046517636718.12026-11-04
Inkling-0.1633213859754221217.32026-11-04
Inkling Small-0.170396345204937215.92026-11-04

常见问题