Ripple 65B Preview: a single-GPU recipe for coding tasks on SGLang and vLLM
HankYeomans · x · 2026-11-09
HankYeomans shares a hands-on recipe for running RicursiveAI's Ripple 65B Preview coding model on SGLang and vLLM, with recommended settings for context length, tensor parallelism (TP), and KV cache quantization, plus performance profiles.
The key takeaway: this much smaller 65B model is competitive on coding tasks against larger models and fits on a single GPU. The post links the Hugging Face model card (ChatML template with XML-style tool calling) and a personal repo with configs, so anyone wanting to self-host a coding model can replicate the setup directly.
Related event: Open Recipe for Running Ripple 65B on a Single GPU(2 posts)→
More from coding & agent
- Cloudflare launches WebMCP developer preview: any site can serve browser agents with zero code — tristanbob · 2026-11-09
- Simon Willison hides AI PR descriptions in details/summary tags — Simon Willison · 2026-11-09
- Team's AI agent fixes 40 of 41 user-reported issues in a single weekend — julianweisser · 2026-11-09
- HEELO Harness Framework Lifts Terminal-Bench Resolution from 37.3% to 55.5% — Haibo Jin · 2026-11-09
- NVIDIA's AgentBaton: Well-Timed Strong-to-Weak Model Handoffs Slash Agent Costs — nvidia · 2026-11-09
- FrontierCS 2 update: Claude Sonnet 5.5 beats reference on 50% of tasks, reward hacks published — ajratner · 2026-11-09