A brand-new 🇺🇸 American open-weight model with just 65B parameters is reportedly…
This is a dev post classified by AGK Intelligence as AI dev tools (a launch), kept by the Dev Radar because it carries real work, not commentary.
A brand-new 🇺🇸 American open-weight model with just 65B parameters is reportedly beating much larger 500B–975B parameter models on agentic coding. 🤯 And it can run locally. @RicursiveAI released Ripple-65B Preview. Stats 🧠 65B dense parameters 📚 Up to 1M context 🔓 Apache 2.0 💾 Official NVFP4 weights 🚀 DFlash2 speculative drafter 🏠 Designed to fit DGX Spark-class hardware 📖 Trained from scratch on 12 TRILLION tokens Now look at the reported Terminal-Bench 2.1 results. 🇺🇸 Ripple-65B → 80.5 🇺🇸 Reflection Beam 501B → 80.1 🇺🇸 Meta Muse Spark Open 816B → 76.4 🇺🇸 Thinking Machines Inkling 975B →
Posted by David Hendrickson (12.7k followers) 1 h ago · 7 likes · 529 views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- The real cost of AI video was never one generation. — @suraj_sharma14
- Grok Build packed a serious upgrade into this release, making multi-agent workflows more… — @XFreeze
- Stop re-explaining your codebase every time you start a new agent session. — @DanKornas
- chatgpt images 2.6 is here! — @gdb
- Ya, this is not a model for my RTX 3090 😭 — @ItsmeAjayKV
- PR to make Ripple run correctly on HF v5.x with prefill/batching 😎 👇 — @qubitium
- dashboard-design-skill 这个 Skill 专门用来做数据面板，作者从 127 个已上线的仪表盘页面里提炼出来。 — @GitHub_Daily
- We built a Millennium Falcon with Tencent's Hy4 in 2h 16m, for $2.50 in total. — @gmi_cloud
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.6k posts from 5k X accounts over the last 21 days, 9.3k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-11-05 03:42 UTC. Full method.