OpenRSI launches benchmark testing recursive self-improvement on 1k-GPU clusters with 60+ hour agent runs
ChengleiSi · x · 2026-09-24
- OpenRSI released OpenRSI-Index v0.1, an open, neutral benchmark for whether AI can recursively improve itself and push the scientific and intelligence frontier beyond human-designed methods.
- Fully open-source projects are turned into autoresearch environments, with agent trajectories lasting 60+ hours on production-scale 1k-GPU clusters; building v0.1 took 100K+ H100-hours.
- The team also demoed RSI-Anything, an agent that converts an RSI idea into a rigorous benchmark problem.
Related event: OpenRSI Launches First Open Benchmark for Recursive Self-Improvement(4 posts)→
More from coding & agent
- Agents keep silently corrupting Office files, so this dev open-sourced Paper Office to fix it — TheMoonMidas · 2026-09-24
- Together AI open-sources tev1, a decision model finetuned on Qwen3.5 4B with full data recipe — nutlope · 2026-09-24
- Muse for Mac adds computer use: queue jobs and walk away — alexandr_wang · 2026-09-24
- Solo Dev Builds MCP Server So Claude Code Ships APKs Straight to Phones — justabigmilkShake · 2026-09-24
- Seroter Daily #873: The Internet Isn't Ready for the Agentic Wave, MCP Apps, GKE Scale-to-Zero — rseroter · 2026-09-24
- Krea Agent adds custom apps: build creative tools by describing them — angrypenguinPNG · 2026-09-24