Inference
EN
Bowen Ye et al.

Claw-Eval

A transparent real-world autonomous-agent benchmark with 300 human-verified tasks, 2,159 rubric items, and Pass^3 scoring across general, multi-turn, and native multimodal agent tasks.

ModelsModels with a result.42
Top resultBest result on this benchmark.81.4%Ornith-1.5-397B
Top-3 spreadPoints from first to third.3.7 pts
YearYear of release.2026

Result and price

27 of 42 results

Results

42 results
#ModelResultThe score from the source.
01Ornith-1.5-397B81.4%
02LongCat-2.578.3%
03K-EXAONE 2.0Above fitFar above this model's index.77.7%
04Ornith-1.0-397B77.1%
05MiniMax M3Above fitFar above this model's index.74.5%
06dots3-note Preview73.4%
07Ornith-1.5-35B-A3B72.5%
08Qwen3.6-27BAbove fitFar above this model's index.72.4%
09Claude Opus 4.670.4%
10Spark-X2.6-4BAbove fitFar above this model's index.69.9%
11Ornith-1.0-35B69.8%
12Qwen3.6-35B-A3B68.7%
13LFM3-1.7BAbove fitFar above this model's index.68.6%
14Claude Sonnet 4.667.8%
15Step 3.7 Flash67.1%
16Ornith-1.5-9B66.5%
17Qwen3.7 Max65.2%
18LLaDA2.2-flash64.2%
19Muse Spark63.8%
19MiMo-V2.5-Pro63.8%
21Ornith-1.0-9B63.1%
22LFM2.5-2.6BAbove fitFar above this model's index.62.9%
23Qwen3.7 Plus62.7%
24Kimi K2.662.3%
24GLM-5.162.3%
24MiMo-V2.562.3%
27GPT-5.460.3%
28Claude Opus 4.559.6%
29Qwen3.6 Plus58.8%
30Gemini 3.1 Pro57.8%
30MiMo-V2-Pro57.8%
32GLM-557.7%
33LLaDA2.2-miniAbove fitFar above this model's index.57.2%
34Qwen3.5 397B56.8%
35GLM-5-Turbo55.8%
36GLM-5V-Turbo53.8%
37Kimi K2.552.3%
38Gemini 3 Flash49.2%
39MiniMax M2.748.7%
40MiMo-V2-Omni45.2%
41DeepSeek V3.240.2%
42Nemotron 3 Super 100B5.5%