Article Complete
AIM
Mistral AI Releases MathΣtral, New Model for Math Reasoning and Scientific Discovery
Mistral AI Releases MathΣtral, New Model for Math Reasoning and Scientific Discovery
Calculating Credibility

Mistral AI Releases MathΣtral, New Model for Math Reasoning and Scientific Discovery

It scores 56.6% on MATH and 63.47% on MMLU, demonstrating superior reasoning capacities within its size category.

Published Updated

Key takeaways

  • Mistral AI launched MathΣtral, a 7B model focused on advanced mathematical reasoning.
  • Achieve superior performance with MathΣtral, scoring 56.6% on MATH and 63.47% on MMLU.
  • Utilize MathΣtral for complex logical reasoning challenges in STEM fields.
  • Access MathΣtral's weights via HuggingFace for easy integration into research projects.
  • Leverage Mistral's tools for initial exploration and fine-tuning of the model.
5 min

Mistral AI has unveiled MathΣtral, a specialised 7B model designed for advanced mathematical reasoning and scientific exploration. Released under the Apache 2.0 license, MathΣtral pays homage to Archimedes on the occasion of his 2311th anniversary this year.

https://twitter.com/theo_gervet/status/1813227214327927082

MathΣtral is tailored to tackle complex, multi-step logical reasoning challenges in STEM fields. Developed in collaboration with Project Numina, the model inherits capabilities from Mistral 7B, achieving state-of-the-art performance across industry-standard benchmarks. Notably, it scores 56.6% on MATH and 63.47% on MMLU, demonstrating superior reasoning capacities within its size category.

Detailed benchmarks highlight MathΣtral's robust performance improvements with increased inference-time computation. For instance, MathΣtral 7B achieves significant accuracy enhancements, scoring 68.37% on MATH through majority voting and 74.59% with a strong reward model among 64 candidates.

MathΣtral is available for immediate use and adaptation using Mistral AI's tools. Developers can deploy the model through mistral-inference for initial exploration and fine-tune its capabilities with mistral-finetune. The model's weights are accessible via HuggingFace, facilitating straightforward integration into academic and research projects.

By releasing MathΣtral to the scientific community, Mistral AI aims to foster advancements in mathematical problem-solving and support academic endeavors. This initiative underscores Mistral AI's commitment to promoting specialized model architectures and their practical applications in scientific discovery.

Membership

Stay ahead with AI insights.

Unlimited access to premium articles, exclusive analysis, and the stories shaping India's AI economy.

Subscribe now

AIM Americas

1190 Miraloma Way Suite P, Sunnyvale, CA 94085

AIM India

1st Floor, Sakti Statesman, Bellandur, Bengaluru – 560103

© 2026 AIM. All rights reserved.