Devstral 2 123B: benchmark results

Provider: Mistral.

Aggregate Rating 1658 ± 50, rank #574 of 10916 rated models, from 8 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
OrgForge-IT100Verdict F1 (imported)94.4
Sonar LLM Leaderboard - Java - Code Smell Density15.28Code smells per 1,000 lines of code (lower is better)86.7
CVerifBench97.2Total (self-reported)69.2
Sonar LLM Leaderboard - Java - Bug Density1Bugs per 1,000 lines of code (lower is better)26.7
Sonar LLM Leaderboard - Java - Vulnerability Density0.36Vulnerabilities per 1,000 lines of code (lower is better)24
TimeSage-EV77.26Overall (%)20
Sonar LLM Leaderboard - Java - Pass Rate68.76Passing tests (%)13.3
BudHist - 2026-04 Ollama Cloud171Correct answers12.5

Interactive version: theaggregate.ai/model?slug=devstral-2-123b · How It Works · Data refreshed daily, snapshot 2026-11-08.