FreshContextA context lab · Fife
Drawing
02 · Benchmarks
Sheet
2 of 3
Rev
A · 2026-10-01

Model · measured only← Models

qwen3.8-27b-mlx-4bit

4 recorded runs. This model has no written entry yet.

Scored runs
4
Run log entries
4
RecordIdentifier

qwen3.8-27b-mlx-4bit

Every result below is as recorded. Without a written entry there is no stated reason for the test and no verdict.

Recorded runs4 scored · 95% intervals
SuiteScore95% interval PassedRuntime · profileRecorded
contract75.0%61.2–85.1Passed 36/48mlxextraction2026-09-08
groundingunenforced92.0%85.9–95.6Passed 115/125mlxextraction2026-09-07
screening93.3%84.1–97.4Passed 56/60mlxextraction2026-09-08
toolcall100.0%94.0–100.0Passed 60/60mlxextraction2026-09-09

Runs marked unenforced were measured through mlx_lm on a suite that requires constrained decoding. mlx_lm accepts the schema and ignores it, so those numbers do not measure what the suite is designed to measure. They are kept and labelled rather than deleted.

Every run log entry for this model (4)
FinishedSuiteRuntime · profileScoreCoverageGrader · dataset · config
toolcalltoolcall/1mlxextraction100.0%Coverage 60/60g 804e8fc9457649d8d 10af88c170513155c 8886d10bc276895c
screeningscreening/1mlxextraction93.3%Coverage 60/60g 0661b49a10be2d2ad 64c97ca2269c29afc 1b2a598c6ef11638
contractcontract/2mlxextraction75.0%Coverage 48/48g 486601ba23034dd6d 1fe53d30cb5480f6c 2750a525d9018a88
groundinggrounding/1mlxextraction92.0%Coverage 125/125g 30b933dcc6a7e2d4d dcb396faf8b83b0ac 42a3cf291ff302bb