RSI Arena: 8 AI agents get 1,000 GPU-hours each to train a better model live
my_cat_can_code · x · 2026-09-29
RSI Arena opens live at COLM 2026: eight AI agents start from the same base model (Nemotron 3.5 Lightning 30B-A3B), each with $300 of API credit and 1,000 GPU-hours on 64× RTX PRO 6000 across 8 Slurm clusters, and get 144 hours to pick data, define benchmarks, and run experiments autonomously.
- Stage 1 kicks off Sept 29, 5 PM PT, streamed at rsiarena.live with a prize round for predicting the top three
- Stage 2 lets humans judge the models in the arena; agents keep training on that feedback with another 500 GPU-hours
- Checkpoints and data will be open-sourced, testing one concrete step of recursive self-improvement
More from coding & agent
- Generic API-catalog MCP server vs one tool per job: a context-bloat experiment — MandeepKaur9407 · 2026-11-16
- Berkeley paper: LLMs answer with stale values even when the update is in context — ombre-ai · 2026-11-16
- Effective HTML: 4.2k-star agent skills turn wireframes into self-contained interactive prototypes — tom_doerr · 2026-11-16
- skills-gate: one gate skill fronts 148 agent skills at just 389 tokens — Distinct-Pie2389 · 2026-11-16
- Open-weight Karpo-Decision-31B tops Decision Index at 73.9, one forward pass per answer — CodeByPoonam · 2026-11-16
- shadcn boosts @uisdk: one-shot UI rebuild that's instantly E2E-testable by AI — shadcn · 2026-11-16