Ripple-65B set U.S. open-weight SOTA in agentic coding; builders share the recipe
Azaliamirh · x · 2026-11-14
Ricursive Intelligence released Ripple-65B Preview last week: a 65B dense LLM pretrained from scratch on 12 trillion tokens with up to 1M context, setting a new U.S. open-weight SOTA on agentic coding. Its small footprint lets it run on local devices like the NVIDIA DGX Spark; an internal variant powers Ricursive's AI chip design platform.
Together AI will host a live technical session on Tuesday, Nov 17, 11am PT with founding staff member Dan Zhang and Zain Hasan, going behind the build: why dense over MoE, which ablations held up (and didn't), what happened inside the RL runs, and how to run Ripple on Together and integrate it into coding agent workflows.
Related event: Ripple-65B Sets US Open-Source Coding SOTA as Team Details Training(3 posts)→
More from coding & agent
- Dev open-sources React Dynamic Island component islandcn after thousands in AI inference costs — HankYeomans · 2026-11-14
- LangChain launches LangSmith platform: one-click agent deploy with 24/7 auto-monitoring — hwchase17 · 2026-11-14
- Tire Swing launches fully autonomous AI to automate affordable housing compliance — ycombinator · 2026-11-14
- 130K-follower KPI... 'Vibe coding as recreation': 15 minutes of prompting to train a model on his messages — gabriel1 · 2026-11-14
- Google's Antigravity multiagent proves 2013 info-theory conjecture in Lean 4 — ymatias · 2026-11-14
- Read-only MCP setup lets Claude Code audit a Proxmox lab — Worldly_Piglet_7040 · 2026-11-14