Today
-
Add measured Codex reconstruction results with Orichalcum pin-only disclosure
-
Rescore saved Sol, GLM and Nemotron outputs with fidelity v7
-
Publish final Sonnet results and three-colour benchmark figures
Yesterday
-
Publish reconstruction readouts and benchmark leaderboard
-
Publish reconstruction readouts and benchmark leaderboard
-
Publish reconstruction readouts and benchmark leaderboard
-
Publish reconstruction readouts and benchmark leaderboard
-
Retain completed SOL-6.1 results and remove other runs
2 days ago
-
Remove superseded Qwen results and medium-reasoning Sol pilot
-
Show measured efficiency values with incomplete coverage markers
-
Publish reconstruction readouts and benchmark leaderboard
-
Backfill retained Sol usage and pricing-derived inference costs
-
Publish reconstruction readouts and benchmark leaderboard
-
Add paired viewer images, confidence intervals and leaderboard plots
3 days ago
-
Publish SimReadyBench under HCLTech Robotics
-
Add fidelity score diagrams
Add formal, visual and kinematic score diagrams with Geist Sans labels, typeset equations and editable Mermaid sources. -
Use Geist Sans and an animated side-by-side comparison
-
Revise benchmark methodology and scientific illustrations
-
Add orientation-aligned comparison to the evaluation methodology
-
Add benchmark methodology, evaluation figures and reference observations
