agents benchmark
Claw-Eval
Claw-Eval tests real-world agentic task completion across complex multi-step scenarios, evaluating a model's ability to use tools, navigate environments, and complete end-to-end tasks autonomously.
Updated Aug 7, 2026
Published models13
Registry coverage13
MetricScore
EvidenceB
Claw-Eval leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | MA | 80.9% | 100.0% | 13 | C | |
| 2 | ZA | 75.0% | 91.7% | 13 | C | |
| 3 | MI | 74.5% | 83.3% | 13 | C | |
| 4 | TE | 68.5% | 75.0% | 13 | C | |
| 5 | AC | 65.2% | 66.7% | 13 | C | |
| 6 | XI | 64.0% | 58.3% | 13 | C | |
| 7 | XI | 63.2% | 50.0% | 13 | C | |
| 8 | AC | 62.7% | 41.7% | 13 | C | |
| 9 | XI | 61.5% | 33.3% | 13 | C | |
| 10 | AC | 60.6% | 25.0% | 13 | C | |
| 11 | AC | 58.7% | 16.7% | 13 | C | |
| 12 | XI | 54.8% | 8.3% | 13 | C | |
| 13 | AC | 50.0% | 0.0% | 13 | C |
Score distribution
Top published rows on the benchmark's original scale.
Claw-Eval highlights
The top published results on this benchmark's own scale.
What is Claw-Eval?
Definition and scoring fields from the benchmark registry.
Claw-Eval tests real-world agentic task completion across complex multi-step scenarios, evaluating a model's ability to use tools, navigate environments, and complete end-to-end tasks autonomously.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- Claw-Eval
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- claw-eval|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about Claw-Eval.
Which model scores highest on Claw-Eval?
Kimi K2.6 is currently ranked first with 80.9%.
What does Claw-Eval measure?
Claw-Eval tests real-world agentic task completion across complex multi-step scenarios, evaluating a model's ability to use tools, navigate environments, and complete end-to-end tasks autonomously.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
13 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.