Advancing science & engineering with Antigravity’s /teamwork
Google Research announced the external release of Antigravity's /teamwork multiagent systems and related Gemini 4 models, which pair specialized generator agents with independent verification mechanisms to solve complex problems across mathematics, chemistry, systems software, and machine learning engineering. Teams applied these tools to automate long-horizon workflows, resulting in verified formal mathematical proofs, molecule structure discoveries, and optimized infrastructure kernels. Open article →
Notable comments
Reception is mixed: commenters point to concrete multi-domain claimed results and some first-hand experimentation, but raise reliability/workflow roughness, opacity vs competitors, and skepticism about whether the approach is meaningfully different or necessary given similar “dynamic workflow” systems.
HN · wslh FWIW, since Argon 4 was launched I used Antigravity (without the /boost or the /teamwork flags) for "playground research" and had some discoveries in cryptography[1] and speeding up Python, removing the GIL, improving the JIT, and optimizing it for certain libraries (e.g. NetworkX) [2].
HN · mirrokni Here is a short list of example results reported on the post; again we will have dedicated Antigravity posts for some of these.
- Google Antigravity agents tackle 90-year-old math problem and rewrite software
- Frontier science doesn't advance through single-turn AI prompts — it advances through rigorous di...
- 🚨HOLY SHIT. Just weeks after OpenAI announced new AI math discoveries, Google has dropped an INSA...
- Advancing science and engineering with Antigravity's /teamwork
- Today, we are sharing new breakthroughs from our Antigravity’s /teamwork multi-agent systems. Thr...
- Check out our recent results solving open challenges across math/science and engineering with /te...