Leanstral: 6B Active Parameters Hits New SOTA in Theorem Proving
AlbertQJiang · x · 2026-08-07
Leanstral is a series of generalist code-agent models for Lean 4 theorem proving. It has 119B total parameters but only 6B active parameters. Instead of a specialized prover scaffold, Leanstral operates directly within the open-sourced Mistral Vibe code-agent harness and uses no test-time-scaling method beyond context compaction, with performance scaling smoothly as the per-problem token budget grows.
Despite its size, it delivers results rivaling far larger and proprietary systems:
- Saturates the miniF2F benchmark.
- Solves 587/672 problems on PutnamBench.
- Achieves a new state-of-the-art 34% on FATE-X and 43.2% on real-repo FLTEval.
Beyond competition mathematics, it resolves issues in real repositories involving graduate-level math and code verification. An automated pipeline built on it uncovered previously unknown bugs in open-source software. The model is open-sourced under Apache-2.0.
Related event: Leanstral Model Released, Setting New SOTA in Theorem Proving(2 posts)→
More from coding & agent
- Opus 5.5 + Three.js demo stuns with water physics, fish and blossoms — nikola_mr64990 · 2026-10-16
- Grok-built linear regression bot claims $508K PnL in 125 days on Polymarket — Roger_M_Taylor · 2026-10-16
- Karpathy's knowledge engine workflow: Claude Code turns raw notes into a self-updating wiki — nikola_mr64990 · 2026-10-16
- Sierra and Meta unveil Personal Agent Protocol with Zapier as launch partner — sujingshen · 2026-10-16
- ArcKit adds 102 AI-narrated short films for commands across 14 community overlays — mcraddock · 2026-10-16
- Personal agents live or die on chaining tasks, not single-task ability — sujingshen · 2026-10-16