MultiChallenge: provider-reported runs
Tests instruction retention, memory of user information, versioned editing, and coherence over multi-turn conversations.
1 models
Higher is better.
- 1Mercury 3Inception60.35
About this benchmark
- Authors
- Scale AI (2025)
- Source
- MultiChallenge ↗
Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.