💡

Complete Your AI Tool Stack

Leanstral users also rely on these tools to enhance their workflow:

ElevenLabs

Try Free

Ultra-realistic AI voiceovers

Add professional narration to your videos

Free tier, $5/mo+Learn more →

Murf.ai

Try Free

Studio-quality AI voices

Create voiceovers in 120+ voices

Free tier, $19/mo+Learn more →

AdCreative.ai

Try Free

AI-powered ad creatives

Generate marketing visuals in seconds

$29/mo+Learn more →

💰 Affiliate disclosure: We may earn a commission if you sign up through these links at no extra cost to you.

Listed in Coding & Development with 135 other toolsPart of 790+ curated AI tools on AISO
Leanstral logo

Leanstral

Mistral's open-source Lean 4 proof agent — formal verification at low cost

0
freemiumDR 86Free API endpoint at launch. Available in Mistral Vibe (zero-setup). Open weights under Apache 2.0 for self-hosting.View full pricing →

Visit Leanstral

https://mistral.ai/news/leanstral

💰 View Detailed Pricing →Try Leanstral →

About Leanstral

Leanstral is Mistral AI's open-source code agent purpose-built for Lean 4, the proof assistant used for formal verification of mathematics and mission-critical software. Released March 16, 2026, it's a 120B-parameter sparse MoE model with 6B active parameters — designed to operate in realistic formal repositories, not just isolated math competition problems. Apache 2.0 license, free API endpoint, and integrated into Mistral Vibe for zero-setup use.

Key Features

✓120B total parameters, 6B active (sparse MoE) — efficient inference at scale
✓Apache 2.0 open weights — full self-hosting permitted
✓Free API endpoint via Mistral La Plateforme
✓Zero-setup integration in Mistral Vibe via /leanstall command
✓MCP support: trained to maximize performance with lean-lsp-mcp
✓FLTEval benchmark: 26.3 at pass@2 — beats Claude Sonnet 4.6 (23.7) at 1/15th the cost
✓At pass@16, reaches 31.9 — beats Sonnet by 8 points and Haiku by 8.9 points
✓Outperforms Qwen3.5-397B-A17B and Kimi-K2.5-1T-A32B despite far fewer active parameters
✓New FLTEval evaluation suite for real proof engineering (FLT project PRs), not just competition math

Leanstral Pros & Cons

✅ Pros

  • +Best open-source model for Lean 4 formal verification — first purpose-built agent for the task
  • +Extreme cost efficiency: pass@2 beats Claude Sonnet at $36 vs $549 (15× cheaper)
  • +Apache 2.0 license — no usage restrictions for commercial or research use
  • +Sparse MoE architecture keeps inference costs low despite 120B total parameter count
  • +Works in realistic proof engineering repos (FLT project), not just toy math problems
  • +MCP-native: trained for lean-lsp-mcp, pluggable with arbitrary MCPs via Mistral Vibe

⚠️ Cons

  • −Highly specialized: only useful for Lean 4 formal verification — not a general coding model
  • −Claude Opus 4.6 still leads on quality (39.6 vs 31.9) when cost is not a constraint
  • −Lean 4 ecosystem is small — most developers will never have a use case for this model
  • −FLTEval is a new benchmark without broad community adoption yet

Who Is Leanstral Best For?

👤Mathematicians and researchers working in Lean 4 on large formal proof repositories
👤Teams at the frontier of formal software verification needing automated proof generation
👤Universities and research labs that need open-weight models for formal methods coursework
👤Organizations building Lean 4 proof pipelines where API cost at scale matters

Tags

mistrallean4formal verificationproof assistantopen sourcecode agenttheorem provingmoeapache 2.0
🏷️

Is this your tool?

Claim your listing to get a Featured badge, edit your description, and stand out from competitors. All plans include a permanent dofollow backlink to your site.

Claim Now →

Stay updated on Coding & Development tools — join our weekly newsletter

One concise email with fresh launches, trending picks, and featured standouts.

Alternatives to Leanstral

View all Leanstral alternatives →

GitHub Copilot

AI pair programmer for code suggestions

paidCompare Leanstral vs GitHub Copilot →

Claude Opus 4.8

Anthropic's flagship model — stronger coding, agents, and honesty

paidCompare Leanstral vs Claude Opus 4.8 →

Devstral 2

Mistral's SOTA open-weight coding model — 72.2% SWE-bench, free API

freemiumCompare Leanstral vs Devstral 2 →