reasoning benchmark
MEWC
MEWC is a benchmark that evaluates AI model performance on multi-environment web challenges, testing agents' ability to navigate and complete complex tasks across diverse web environments.
Updated Aug 11, 2026
Models1
Model coverage1
MetricScore
EvidenceB
MEWC Ranking
Higher score ranks better on this benchmark.
| 1 | MI | 74.4% | 100.0% | 1 | C |
MEWC Highlights
The leading models and scores on this benchmark.
What is MEWC?
What MEWC measures and how its scores work.
MEWC is a benchmark that evaluates AI model performance on multi-environment web challenges, testing agents' ability to navigate and complete complex tasks across diverse web environments.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MEWC
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- mewc|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MEWC.
Which model scores highest on MEWC?
MiniMax M2.5 is currently ranked first with 74.4%.
What does MEWC measure?
MEWC is a benchmark that evaluates AI model performance on multi-environment web challenges, testing agents' ability to navigate and complete complex tasks across diverse web environments.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 model results are currently shown.
Does this benchmark affect the overall score?
No. This benchmark is shown for reference but does not contribute to the overall score.