Fetching from the wire…
01
02
03
04
05
06
07
08
OSS2026-08-15 · source-backed
The mini_tool_call suite was effectively ranking brace hygiene, so Llama-3.1-8B picked the correct tool every time yet scored 0.225 because it emitted three opening braces and two closing ones and the scorer rejected the bounded-scan result. GitHub Every defect was reproduced against shipped v0.73.1 before a line changed, using a stub emitting real model output shapes instead of a GPU. More eval harnesses should adopt that discipline. Soup's separate headline claim, fine-tuning an 8B model on a 4GB laptop GPU from one YAML file, rests on the project's own description with no independent benchmark, so treat it as unverified.
Each link below shares sources, entities, or timing with this story.
Meta released Llama / Shared entity: GitHub / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Meta released Llama); both cover GitHub; reported by the same outlet (github.com).
Meta released Llama / Shared entity: GitHub / Same source domain / Earlier coverage
Linked by a graph relationship (Meta released Llama); both cover GitHub; reported by the same outlet (github.com).
Meta released Llama / Shared entity: GitHub / Shared topic / Earlier coverage
Linked by a graph relationship (Meta released Llama); both cover GitHub; overlapping topics (against, model).
Meta released Llama / Shared entity: Llama / Shared topic / Earlier coverage
Linked by a graph relationship (Meta released Llama); both cover Llama; overlapping topics (benchmark, model).
Linked by a graph relationship (Meta released Llama); both cover Llama; overlapping topics (benchmark, model).
Linked by a graph relationship (Meta released Llama); both cover Llama; overlapping topics (benchmark, model).
Meta released Llama / Shared entity: GitHub / Same source domain / Earlier coverage
Linked by a graph relationship (Meta released Llama); both cover GitHub; reported by the same outlet (github.com).
Meta released Llama / Shared entity: GPU / Earlier coverage / Tension
Linked by a graph relationship (Meta released Llama); both cover GPU; earlier GPU coverage from 2026-07-25.