Laguna M.1 — benchmark results
Poolside's open 225B MoE (23B active) agentic coding model trained with code-execution RL, scoring 72.5% on SWE-bench Verified (April 2026). Provider: Poolside. Released 2026-04-28. Access: Open.
Unified ELO 1498 ± 45, rank #806 of 1776 rated models, from 26 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Wolfram LLM Benchmarking Project | 46 | Correct Functionality (%) | 61.8 |
| Vals AI CorpFin v2 | 58.16 | Accuracy (%) | 37.7 |
| Vals AI Terminal-Bench 2.0 | 31.46 | Accuracy (%) | 36.4 |
| SvelteBench | 80 | Average pass@1 (%) | 31.4 |
| Gert Labs Rankings | 38.87 | GScore (%) | 30.2 |
| Chatbot Arena (Code) | 1352 | Elo | 29.3 |
| Vals AI LiveCodeBench | 61.44 | Accuracy (%) | 29.1 |
| Vals AI LegalBench | 75.14 | Accuracy (%) | 24.8 |
| Vals AI Vibe Code Bench | 11.04 | Accuracy (%) | 18.6 |
| Vals AI ProgramBench | 19.76 | Raw Pass Rate (%) | 16.7 |
| PM-LLM-Benchmark | 25.2 | Score | 15.7 |
| SpeechMap Compliance | 38.2 | % Requests Completed | 15.2 |
Interactive version: theaggregate.ai/model?slug=laguna-m-1 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.