FrontierCS 2 launches on Kaggle with DeepMind: 48 tasks ask agents to beat published papers
lmoroney · x · 2026-10-27
FrontierCS 2 is live on Kaggle, co-launched with Google DeepMind. Built from published papers across 13 CS areas, its 48 tasks delete each paper's contribution from the released code and challenge agents to beat it on the authors' own metrics — no paper, no internet, five hours. The design doubles as a training exercise for research engineers: can you rediscover an idea from code and metric alone? Kaggle says plenty of headroom remains.
Related event: DeepMind and Kaggle Launch FrontierCS 2 Benchmark(2 posts)→
More from coding & agent
- Batch-editing hundreds of drone videos by scripting DaVinci Resolve with Astra — sethlazar · 2026-10-27
- DeepSeek V4.1 Flash lands in Cursor, dominating the cost frontier at $0.60-$0.98 per task — dean_rie · 2026-10-27
- Masking slides for AI: placeholders that survive the model but break the layout — shoumikgoswami · 2026-10-27
- Creator has Claude edit her iPhone Duo review inside Tella's agent video editor — evielync · 2026-10-27
- Ant International and Westlake University release Trajectory2Gym for tool-use agent training — ant-intl · 2026-10-27
- Eight universities propose unified framework for work agents — Zishan Xu · 2026-10-27