FrontiersMind open-sources Lumma-Fev decision models from 154M to 9B under Apache 2.0
kalyan_kpl · x · 2026-09-29
FrontiersMind released Lumma-Fev, a family of decision models that score user-supplied options in a single forward pass — no autoregressive decoding or JSON repair — for tasks like ticket routing and urgency ranking. Four sizes (154M, 649M, 4B, 9B) ship under Apache 2.0, with a blog covering the API, outputs, and failure modes.
More from Models
- Claude usage limits loosened dramatically: devs report going from barely usable to unlimited — WhatTheLJW · 2026-09-29
- Cekura speech benchmark: 2,214 live calls, GPT Realtime 2.1 most reliable, Phonic v1 fastest at 1.59s — graceisford · 2026-09-29
- User: Opus 5.5 powering OpenClaw agent already beats months-old GPT-5.6 setup — heyneighbor · 2026-09-29
- Bittensor Subnet 120 launches Affine-1, a 36B decentralized model claiming SOTA — markjeffrey · 2026-09-29
- Claude Design's Capabilities Are Baked into Opus 5.5, Argues AI Observer — dotey · 2026-09-29
- GPT-6 Sol appears on LMArena: 24-hour Direct Mode window before Battle and Agent Mode — arena · 2026-09-28