Local LLM shootout: Flint 40B MoE vs Qwen3.8-27B on identical RTX A6000s
RichardSocher · x · 2026-10-23
Recursive's Flint model is getting side-by-side tests from the local LLM community. User nimorotem ran Flint vs Qwen3.8-27B at 8-bit on two identical RTX A6000 48GB cards with a live comparison site (4-bit is even faster) and says he's switched from being a heavy Qwen user to 100% Flint; jtdavies counters that the 40B Flint should be compared to similar-size models. Setup: Flint is a 39.8B-total/2.5B-active MoE with 262K context on llama.cpp, Qwen a 27B dense model on vLLM; identical sampling (temp 1.0, top-p 0.95, top-k 20, 8192 reasoning tokens).
More from Infra
- vLLM details M3.1 Flash serving: NVFP4 cuts weights 44.6%, boosts capacity 2.7x — vllm_project · 2026-10-23
- M3.1 Flash hits 191.5k tok/s per GPU on Vera Rubin NVL72, leads open models on agent benchmarks — vllm_project · 2026-10-23
- Agents that don't close connections on retry waste ~148 tokens per task, vLLM tests show — No-Peanut-6988 · 2026-10-23
- Ex-Googler with 35 Years in Software Pursues Local AI Coding and Music Video Generation — julianharris · 2026-10-23
- Intelligence too cheap to meter: Opus price down 4x while AA score jumps 21 to 58 — haider1 · 2026-10-23
- SK Hyper in talks to fund envisioned $350B AI data center buildout in South Korea — dinabass · 2026-10-23