Mistral AI has released Leanstral 1.5, a 7-billion parameter model specialized in generating proofs for the Lean theorem prover. Building on the previous version, it significantly boosts proof completion rates and offers a more accessible way for developers and mathematicians to use automated reasoning.
Background
- Mistral AI is a French company that builds large language models (LLMs) competing with OpenAI, Google, and Meta. Their models are known for being open-weight and efficient.
- Leanstral is a specialized version of Mistral's models fine-tuned to generate formal proofs for the Lean theorem prover — a software tool that checks mathematical reasoning at the level of each logical step.
- Formal proof systems like Lean are used to verify critical software (e.g., operating systems, compilers) and to give mathematicians absolute certainty about proofs too complex to check by hand. However, writing proofs in Lean is slow and requires rare expertise.
- This release (Leanstral 1.5) claims to produce dramatically more correct proofs than its predecessor, bringing AI-assisted formal verification closer to practical, everyday use. The "proof abundance" in the title refers to making formal proofs plentiful and accessible rather than scarce and expensive.
Max Weinbach says he had early access to OpenAI's new model GPT-5.6 Sol, calling it his favorite model by far. He highlights that it never gives up and will keep reasoning until it's done. OpenAI announced that GPT-5.6 Sol, along with Terra and Luna, will launch publicly on Thursday, with preview access expanding globally now.
The US government ordered Anthropic to suspend access to its Fable 5 and Mythos 5 models for all customers, citing a potential jailbreak technique that involved asking the model to review a codebase for vulnerabilities—a capability Anthropic says is available in other public models. Access was abruptly cut off on June 12.
Andrej Karpathy announces the release of Claude Fable 5, the same underlying model as Mythos but with added safeguards. He calls it a major step forward, particularly for long problem-solving sessions on difficult tasks, and describes it as state-of-the-art on nearly all benchmarks with exceptional performance in software engineering, research, and vision.
Roman Storm warns that the legal theory in his case could set a precedent making open-source developers liable for how others use their code, potentially criminalizing the mere publication of privacy, messaging, or crypto tools. He notes that developer Michael Lewellen cannot publish lawful code due to prosecution fears, and argues this chilling effect extends beyond any single case.
Meta's engineering culture is deteriorating under Mark Zuckerberg and Scale AI CEO Alexandr Wang, who have introduced keyboard tracking, reassignments to data labeling, and AI-centric performance metrics. Critics argue this incentivizes performative AI use, drives away experienced engineers, and contributed to a major Instagram hijacking incident caused by AI-written and AI-reviewed code.