Theorem Proving Costs Begin to Rise: Mistral Releases Open Source Leanstral 1.5, Approximately $4 Per Problem

According to monitoring by Dongcha Beating, Mistral AI has released Leanstral 1.5, a model designed for formal proofs in Lean 4. The model has a total of 119 billion parameters, with approximately 6.5 billion active parameters, and is licensed under the Apache-2.0 protocol, offering free API access. Official evaluations show that Leanstral 1.5 solved 587 out of 672 problems on the PutnamBench; it achieved 87% and 34% on the abstract algebra benchmarks FATE-H and FATE-X, respectively, setting new performance records among similar models. The average cost per problem for Leanstral 1.5 on PutnamBench is about $4, significantly lower than the costs of several previous systems, which ranged from tens to hundreds of dollars. As the token budget per problem increases, the number of problems it can solve continues to rise; in the complexity proof of AVL trees, the model completed the relevant proof after reasoning over more than 2.7 million tokens and 22 context compressions. In addition to mathematical proofs, Leanstral 1.5 has also been used for code verification. The team discovered 11 real bugs in 57 open-source Rust repositories, five of which had not been previously reported.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
  • Reward
  • Comment
  • Repost
  • Share
Comment
Add a comment
Add a comment
No comments
  • Pinned