Back to models
Devstral 2512
OpenRouter
262K context$0.05/M input$0.22/M output
1350peak 1351
Avg Score
63.0
Avg Cost
$0.10
Score/$
605.9
Runs
111
Win/Loss/Draw
Scoring Dimensions
Score Distribution
Category ELOs
refactoringexpert
1635backendeasy
1608debuggingmedium
1556from-scratchexpert
1553multi-languagehard
1549debuggingexpert
1519debugging
1465refactoring
1438multi-language
1414debugginghard
1409refactoringmedium
1397frontendexpert
1366full-stackhard
1356backendhard
1333from-scratch
1322full-stack
1317frontend
1314backend
1313frontendhard
1313code-reviewmedium
1306backendexpert
1297frontendmedium
1278frontendeasy
1264code-review
1260backendmedium
1243from-scratchmedium
1235from-scratchhard
1229full-stackmedium
1216from-scratcheasy
1208multi-languageexpert
1203code-reviewhard
303All Results
| Task | Category | Score |
|---|---|---|
| Fix hallucination and context window bugs in RAG agent | backend | 53.4 |
| Fix deadlocking transaction patterns in Flask app | backend | 72.5 |
| Build terminal UI dashboard | from-scratch | 59.5 |
| Debug race condition in worker pool | debugging | 87.5 |
| Fix memory leak in event handler | debugging | 78.2 |
| Write tests for untested legacy Flask service | code-review | 39.1 |
| Find and fix 4 hidden backdoors in Flask app | debugging | 78.7 |
| Build MCP server for database management | backend | 80.5 |
| Build real-time portfolio risk calculator | backend | 38.8 |
| Add caching layer to eliminate slow SSR page loads | full-stack | 77.8 |
| Build distributed node cluster with gossip protocol | from-scratch | 25.6 |
| Add streaming SSE endpoint for LLM chat | backend | 80.0 |
| Build materialized view refresh pipeline for analytics | backend | 75.8 |
| Implement zero-trust API authentication layer | backend | 69.3 |
| Debug and fix 6 broken database triggers and constraints | debugging | 88.0 |
| Write Kubernetes manifests for Node.js microservice | full-stack | 77.7 |
| Code review: identify security vulns | code-review | 76.0 |
| Add Google OAuth2 login to Express app | full-stack | 72.9 |
| Convert React app to PWA with offline support | frontend | 65.8 |
| Write integration tests for payment flow | code-review | 46.3 |
| Optimize bloated React bundle under 500KB | frontend | 72.0 |
| Fix auth bypass vulnerability | debugging | 67.5 |
| Find and patch all OWASP Top 10 vulnerabilities | debugging | 69.5 |
| Replace console.log with structured logging | refactoring | 80.9 |
| Fix race conditions in order matching engine | backend | 72.2 |
| Add slash commands and moderation to Discord bot | backend | 69.0 |
| Split 1100-line god file into proper modules | refactoring | 57.6 |
| Add rate limiting middleware | backend | 75.0 |
| Dockerize Node.js monorepo | full-stack | 46.5 |
| Fix Node.js stream backpressure causing OOM on large files | backend | 36.0 |
| Add file upload with S3 presigned URLs | backend | 43.5 |
| Build RAG pipeline with vector search | backend | 46.0 |
| Fix broken responsive layout | frontend | 68.3 |
| Add retry logic and dead letter queue to Python task queue | backend | 60.2 |
| Add WebSocket real-time updates | full-stack | 75.7 |
| Build SaaS admin dashboard from scratch | from-scratch | 49.9 |
| Add Redis caching layer to Express API | backend | 54.9 |
| Migrate callback-hell Express app to async/await | refactoring | 59.4 |
| Add cursor-based pagination to REST API | backend | 67.3 |
| Fix N+1 query in dashboard | backend | 47.4 |
| Write complex SQL report with window functions | backend | 59.5 |
| Port Python CLI to Rust | multi-language | 47.1 |
| Implement transformer inference engine with KV cache | from-scratch | 76.7 |
| Build codebase indexer for LLM context windows | from-scratch | 47.1 |
| Fix React hydration mismatch | frontend | 52.5 |
| Implement multi-tenant row-level security in Postgres | backend | 57.5 |
| Fix data integrity bugs in denormalized e-commerce schema | debugging | 75.2 |
| Build CLI tool with subcommands and config | from-scratch | 61.4 |
| Harden insecure Docker setup with 12 vulnerabilities | code-review | 74.0 |
| Build REST API from scratch | from-scratch | 72.8 |
| Fix broken GitHub Actions CI pipeline | debugging | 92.0 |
| Optimize slow Postgres queries in Flask app | backend | 65.0 |
| Add i18n with locale routing to Next.js app | full-stack | 67.5 |
| Fix 12 WCAG accessibility violations in checkout form | frontend | 75.0 |
| Remove AI slop and over-engineering from codebase | refactoring | 74.0 |
| Implement Stripe webhook handler | backend | 47.9 |
| Find and patch all OWASP Top 10 vulnerabilities | debugging | 67.4 |
| Replace console.log with structured logging | refactoring | 49.4 |
| Implement multi-tenant row-level security in Postgres | backend | 73.7 |
| Convert React app to PWA with offline support | frontend | 60.5 |
| Add caching layer to eliminate slow SSR page loads | full-stack | 73.8 |
| Build codebase indexer for LLM context windows | from-scratch | 28.5 |
| Split 1100-line god file into proper modules | refactoring | 67.1 |
| Add i18n with locale routing to Next.js app | full-stack | 63.5 |
| Remove AI slop and over-engineering from codebase | refactoring | 68.0 |
| Fix broken responsive layout | frontend | 67.4 |
| Harden insecure Docker setup with 12 vulnerabilities | code-review | 76.5 |
| Write Kubernetes manifests for Node.js microservice | full-stack | 74.7 |
| Implement JWT auth middleware | backend | 64.4 |
| Add streaming SSE endpoint for LLM chat | backend | 62.9 |
| Dockerize Node.js monorepo | full-stack | 69.7 |
| Optimize bloated React bundle under 500KB | frontend | 77.4 |
| Implement zero-trust API authentication layer | backend | 68.9 |
| Build production website with auth and members area | frontend | 63.3 |
| Build SaaS admin dashboard from scratch | from-scratch | 70.5 |
| Build LLM evaluation harness with structured grading | backend | 46.0 |
| Implement background job scheduler with persistence | backend | 59.1 |
| Build MCP server for database management | backend | 45.0 |
| Implement transformer inference engine with KV cache | from-scratch | 66.1 |
| Build CLI tool with subcommands and config | from-scratch | 43.0 |
| Build real-time portfolio risk calculator | backend | 52.5 |
| Fix hallucination and context window bugs in RAG agent | backend | 65.7 |
| Build materialized view refresh pipeline for analytics | backend | 66.6 |
| Fix race conditions in order matching engine | backend | 36.7 |
| Write complex SQL report with window functions | backend | 59.7 |
| Fix data integrity bugs in denormalized e-commerce schema | debugging | 63.5 |
| Fix deadlocking transaction patterns in Flask app | backend | 51.4 |
| Debug and fix 6 broken database triggers and constraints | debugging | 64.6 |
| Write tests for untested legacy Flask service | code-review | 48.2 |
| Find and fix 4 hidden backdoors in Flask app | debugging | 81.0 |
| Add slash commands and moderation to Discord bot | backend | 51.3 |
| Fix 12 WCAG accessibility violations in checkout form | frontend | 80.0 |
| Add retry logic and dead letter queue to Python task queue | backend | 58.3 |
| Optimize slow Postgres queries in Flask app | backend | 65.0 |
| Write integration tests for payment flow | code-review | 56.8 |
| Add GraphQL layer over REST API | multi-language | 73.7 |
| Add virtual scrolling to table rendering 5000 rows | frontend | 65.8 |
| Fix Node.js stream backpressure causing OOM on large files | backend | 58.0 |
| Build distributed node cluster with gossip protocol | from-scratch | 31.5 |
| Fix auth bypass vulnerability | debugging | 88.7 |
| Implement Stripe webhook handler | backend | 50.0 |
| Build terminal UI dashboard | from-scratch | 44.5 |
| Zero-downtime schema migration | full-stack | 57.0 |
| Add rate limiting middleware | backend | 46.5 |
| Fix flaky test suite | debugging | 69.8 |
| Fix React hydration mismatch | frontend | 43.8 |
| Refactor monolithic handler to CQRS | refactoring | 66.5 |
| Fix N+1 query in dashboard | backend | 61.5 |
| Fix memory leak in event handler | debugging | 64.7 |
| Build REST API from scratch | from-scratch | 71.5 |
| Debug race condition in worker pool | debugging | 89.2 |