# Leanstral by Mistral

15 Hacker News commenters recommended Leanstral by Mistral and 12 criticized it, November 2025 to November 2026. It was recommended in 2 of 13 months.

**Verdict:** Open-weight Lean proof model people can run locally on consumer hardware. Main caveat: critics find it trails larger models on coding and proof tasks.

## Why people recommend it

- Open weights, runs locally (3 comments): "open weights + relatively small is still headline worthy. This thing really cooks." ([source](https://news.ycombinator.com/item?id=48782094)) "We can run Leanstral locally on commodity/consumer hardware." ([source](https://news.ycombinator.com/item?id=48782261)) "It looks like the weights are Apache-2.0 licensed." ([source](https://news.ycombinator.com/item?id=47406282))
- Useful addition to proof toolkits (3 comments): "would be a powerful complement to data‑driven model discovery" ([source](https://news.ycombinator.com/item?id=47416547)) "Try out Leanstral 1.5 on the latest version of OpenATP!" ([source](https://news.ycombinator.com/item?id=48781603)) "this sounds like a great tool to add to the toolbelt" ([source](https://news.ycombinator.com/item?id=48781892))

## What people warn about

- Trails larger models on coding (3 comments): "This model is specifically trained on this task and significantly[1] underperforms opus." ([source](https://news.ycombinator.com/item?id=47405507)) "I found it worse than gemma4 at coding even though it's 4x the parameter size" ([source](https://news.ycombinator.com/item?id=48741616)) "Haiku kinda sucks at this task, and Leanstral is worse?" ([source](https://news.ycombinator.com/item?id=47405328))
- Benchmark examples seem cherry-picked (3 comments): "I don't think I'd consider that such a smashing success that it's worth bringing up as the sole example tbh" ([source](https://news.ycombinator.com/item?id=48781799)) "This is nice work, but I found the bug finding example to be weird" ([source](https://news.ycombinator.com/item?id=48781193)) "something felt cherry-picked" ([source](https://news.ycombinator.com/item?id=48747861))
- Weak docs, FAQ and marketing (4 comments): "Non-competitive products and models combined with bad marketing and GTM" ([source](https://news.ycombinator.com/item?id=48743347)) "just a half-assed FAQ, that seems vibe-coded and searched poorly" ([source](https://news.ycombinator.com/item?id=48742312)) "I'm amazed at how they manage to fumble releases over and over" ([source](https://news.ycombinator.com/item?id=48742572))
- Access restricted behind GitHub or Labs (2 comments): "enabling Labs models isn't available for self-serve activation on standard individual accounts" ([source](https://news.ycombinator.com/item?id=48745202)) "I must connect to GitHub to use" ([source](https://news.ycombinator.com/item?id=48744870))

## Alternatives hackers recommend

- [DeepSeekMath-V2](https://hackerapproved.com/p/deepseekmath-v2.md): 10 recommend
- [Harmonic Aristotle](https://hackerapproved.com/p/aristotle.md): 10 recommend

## Cite as

Hacker Approved, "Leanstral by Mistral", https://hackerapproved.com/p/leanstral-1-5/ (counts from 1,795,640 Hacker News comments, November 2025 to November 2026).