Recursive SI's Flint: local coding model built in 4 months, 125+ tokens/sec on MacBook Pro
rohanpaul_ai · x · 2026-10-22
Richard Socher announced Flint, Recursive SI's first model: a fast local coding model built from scratch in under 4 months by their automated research system, the Eureka Machine.
Key points:
- Claims best-in-class fast local coding: 125+ tokens/sec on MacBook Pro, 150+ on RTX Spark — up to 2x faster than comparable models
- Open weights available on Hugging Face; private, unmetered intelligence running on your own machine
- Notable as a showcase of an automated research system directly producing a model
More from coding & agent
- Nous Research ships official Hermes Agent plugin for Meshy AI 3D generation via MCP — NousResearch · 2026-10-22
- Support AI Regurgitating Stale Docs Is a Growing Failure Mode — Give It Real Platform View — joshwhiton · 2026-10-22
- CallBaba launches BabaBench, a benchmark testing AI agents on real patient advocacy tasks — mattturck · 2026-10-22
- 70%-95% failure rates: devs say AI agents still too unreliable for real workflows — Luke17man · 2026-10-22
- How do you actually test agents before production? Devs tackle failure-path testing — Last-Bluejay-4443 · 2026-10-22
- A Reddit long-read proposes the Risk and Judgement pattern for bounding agent autonomy — mostly_deterministic · 2026-10-22