Recursive ships Flint: 40B MoE with 2.5B active params for laptop coding at 125 tok/s
_rockt · x · 2026-10-22
Recursive introduced Flint, a mixture-of-experts model for fast agentic coding on local hardware: 40B total parameters with just 2.5B active per token, built from scratch in under four months by a small team plus agents. It leads sub-50B MoE models on SWE-Bench Pro v2, Terminal-Bench 2.1 and SWE-Bench Verified, rivals 25–40B dense models while running 5x faster, and hits 125+ tok/s on an M4 Pro MacBook Pro and 150+ tok/s on a 64GB RTX Spark. Flint doubles as a proof of Recursive's Eureka Machine, its automated AI R&D system that developed nearly every part of the stack.
More from Infra
- SPCX earnings call highlights 'computer per watt' metric as AI value shifts down the stack — mobileraj · 2026-10-22
- Eureka Machine runs on Mac inference so quietly the author's MacBook barely spins its fans — cong_ml · 2026-10-22
- Compute demand is so hot it's becoming a month-to-month auction, analysts say long-term contracts are dying — BenBajarin · 2026-10-22
- SGLang ran early RL training on TPUs with Miles and TorchTitan, with promising results — ying11231 · 2026-10-22
- CXMT's market cap tops all mid-size Asian fabs combined, half of SK Hynix — pstAsiatech · 2026-10-22
- SemiAnalysis: DeepSeek's open-sourced DeepInfer lacks key inference-serving components — cHHillee · 2026-10-22