Physics-IQ / Verified / Image / Evidence

Odyssey-3

+ best-of-8 (WMReward + MBR consensus, rank-sum) (custom prompts) / i2v · Not specified by source

Current report

Reported scores

Physics-IQ score: 52.83%

Read the original leaderboard

Provenance

Result origin: Published by Physics-IQ; not run by Rather
Evaluation source: Evaluator unspecified
Source publication date: Not reported
Imported: Oct 13, 9:33 AM UTC
Protocol: Verified / Image
Checkpoint: Not pinned by source

Generated by: See linked submission report
Scored by: See linked submission report
Score as reported: 52.83 🥉 i2v
Source-listed date: 2026-10-08

Submission report

Original and Verified scores use different evaluation methods. Image and multiframe conditioning are separate protocols. Prompting, best-of-N selection and uncertainty remain as reported; leaderboard inclusion does not establish independent organizer evaluation.

Source revision & integrity

Upstream revision: c2bca8a5bf452d24e177e9975641af1888eb1492
Snapshot SHA-256: ca775072cdfa8f9c976e30dd8794f9dce17056c05238086c3306a1b2b6af730c
Original label: Odyssey-3 + best-of-8 (WMReward + MBR consensus, rank-sum) (custom prompts) | i2v

Loading benchmark results…

Odyssey-3

Reported scores

Provenance