Add cagliostro-v1 (157M) on bench 7-2026

Fix missing comma after PixelModel v6 entry (broke JSON parsing)

Update models.json

Add qwen-qwen3.5-9b (bench-effortless-7-2026, bench-easy-7-2026, bench-mid-7-2026) — submitted by ereniko, approved by wop

Add conversate v2 base (#11)

Upload models.json (#10)

Update models.json

Upload models.json

Update models.json

Upload models.json

fix T2I: single ShellD, add FWKV-Image, remove nuShellD/dupes

restore models.json to original, fix ShellD with trusted benchmark

fix: add PixelModel v5, dedup T2I models, add charts

Upload models.json

Update models.json

Update models.json

Added `AtomixLabs/Photon-1.0-1M` (#8)

Update models.json

7-2026 upgrade: models.json

Submitting LFM2.5 models results. (#7)

Update models.json

Update models.json

add instruct qwen 2.5 0.5B

fix mistake (#6)

Re-eval with official script (#5)

Update models.json

Re-eval with official script

v2: fix easy/mid metric labels, update row counts, flag pre-v2 scores stale

Re-evaluation of Qwen 0.5B (#3)

Update models.json

Update models.json

Update models.json

Upload 5 files

Upload 5 files

Upload 6 files