MEWC
Progress Over Time
Interactive timeline showing model performance evolution on MEWC
State-of-the-art frontier
Open
Proprietary
MEWC Leaderboard
1 models
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  MiniMax | 230B | 1.0M | $0.30 / $1.20 |
Notice missing or incorrect data?
What is 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.
MEWC is a text benchmark evaluating models on reasoning and agents tasks. LLM Stats tracks 1 models on this benchmark, scored on a 0–1 scale. The current average is 0.7, with the leader at 0.7.
Compare leaders on the best AI for reasoning and best AI for agents leaderboards.
Current leaders
MiniMax M2.5 from MiniMax currently leads the MEWC leaderboard with a score of 0.744 across 1 evaluated AI models.
FAQ
Common questions about the MEWC benchmark and leaderboard.