Very excited to present Minerva🦉: a language model capable of solving mathematical questions using step-by-step natural language reasoning. Combining scale, data and others dramatically improves performance on the STEM benchmarks MATH and MMLU-STEM. goo.gle/3yGpTN7

Super excited to share Minerva!! – a language model that is capable of solving MATH with 50% success rate, which was predicted to happen in 2025 by Steinhardt et. al. (bounded-regret.ghost.io/ai-forecasting)! #Minerva 1/

Thrilled to announce🦉Minerva: a large language model capable of solving mathematical problems using step-by-step reasoning in natural language. See blog here: goo.gle/3yGpTN7 and samples here: minerva-demo.github.io (1/n)

Very excited to announce a significant milestone in expanding reasoning capabilities of language models! 🎉🎉 We introduce #Minerva🦉: a language model that can solve mathematical questions using step-by-step natural language reasoning: bit.ly/3OBj2d5 🧵 1/

1/ Super excited to introduce #Minerva 🦉(goo.gle/3yGpTN7). Minerva was trained on math and science found on the web and can solve many multi-step quantitative reasoning problems.

Happy to release our work on Language Model Cascades. Read on to learn how we can unify existing methods for interacting models (scratchpad/chain of thought, verifiers, tool-use, …) in the language of probabilistic programming. paper: arxiv.org/abs/2207.10342

AI coding tools are everywhere. Yet they all fall short when scaling to 100s of developers and complex codebases. We’re solving this problem Augment Code. Introducing Augment Code, the first developer AI purpose-built for teams.🧵

Working on open source is hard. We just made it a little easier. Augment is free for open source contributors and maintainers. 🎉 Augment’s deep codebase understanding tears down some of the biggest pains for working on open-source projects– the context you need to

Happy to share that Augment is now available to open source developers, for free! Experience powerful codebase awareness baked into every single feature -- it's a game changer.

Just bumped into John Yang on Stanford campus and discussed future of coding agents and swebench/sweagent. Very happy such smart and kind folks are working on this important problem with an open-source POV. The future is bright ☀️ Strongly believe open source is critical

Augment’s agent is #1 on SWE-bench!

Augment Agent is now live. Designed for large-scale codebases — with 200K context tokens, persistent memory, and deep tool integrations, Augment Agent is available in VS Code and JetBrains.

Been using Augment's Agent continuously, and its deep codebase understanding is unmatched. I go from ticket, through design and implementation, to PR and review, all without leaving Agent mode. Haven't had to write a line of code myself in a while, which is a surreal experience.

Daunting task of the day: Add a feature to our frontend, which I'm not familiar with. Augment's Agent can answer deep questions like "explain what happens after the user clicks Send" and help me cut through the complexity. Augment Code ftw

Introducing Remote Agent - cloud-based dev agents with full-codebase context, deep IDE integration, and full toolchain access. Available in VS Code.

Excited to announce Augment Remote Agents! Once you get used to coding with agents, it becomes obvious that one is not enough. Coding with 5 agents working on parallel on different tasks feels like a superpower. augment.new

I will be going on @TBPN at 1:40pm PT to talk about GPT-5 + Augment Code and share our early impressions. Make sure to tune in!

Great conversation with @matthewberman and Hiten Shah today! Talked about why context is king in AI coding and our unlimited context approach at Augment Code. Hope other folks get access to the model I'm using as my daily driver soon... 👀 Watch it here: youtube.com/live/e7Wmnslt6…