Ricursive Intelligence launches Ripple-65B, claiming open-weight agentic coding SOTA
xyzrt · reddit · 2026-11-05
Ricursive Intelligence released Ripple-65B Preview, a 65B dense model trained on 12T tokens with 1M context, claiming a new US open-weight SOTA on agentic coding: 80.5% on TerminalBench 2.1, 79.1% on SWE-Bench Pro v2.0-Hard, 44.3% on DeepSWE v1.1 — reportedly beating far larger models. Unverified by third parties.
Related event: Ricursive Open-Sources Ripple-65B, Claims US Agentic Coding SOTA(16 posts)→
More from Models
- Ripple-65B claims open-weight agentic coding SOTA trained from scratch for just $32M — schwentker · 2026-11-05
- Ripple claims frontier-level open-weight model trained in 32 minutes at one-tenth cost — annadgoldie · 2026-11-05
- Developers report Grok API instability: 13.58% error rate and 13.6s p90 latency in dashboards — ChrisUniverse · 2026-11-05
- Ramp's post-trained Qwen beats Opus on finance QA; 10,000-agent swarm finds 7 critical bugs — AI Engineer · 2026-11-05
- Vals AI: four of seven frontier models keep 80%+ scores at lowest reasoning effort for half the cost — ZacharyHuang12 · 2026-11-05
- Sailing sim race shows Chinese models lag frontier: "cheap tokens ≠ cheap agents" — shekitup · 2026-11-05