organizational research article

Summary of METR's predeployment evaluation of GPT-5.6 Sol

METR · Published 2026-06-26

Read the original document ↗

Evidence from this source

Results and research linked to this document
Type	Evidence
Result	Task-completion time horizons · GPT-5.6 Sol / METR horizon evaluation Source details 50%-Time Horizon paragraph, cheating attempts scored as failures

Document tracking details

Source ID: src-metr-gpt56-sol
Source type: Primary
Original URL: https://metr.org/blog/2026-06-26-gpt-5-6-sol/
Last updated by publisher: Not reported
Access status: available
Retrieved: 2026-09-27T04:17:28Z
Document fingerprint: 232140ff97a4f324e2b1bdd918d243be6f0a16ddc999ca0e3d90d7b6ece002bf
Based on original bytes. Used to detect changes to the source.

Availability checks

2026-09-27T04:17:28Z · success: Original public METR assessment inspected for 50%-horizon point and interval.
2026-09-27T05:02:16Z · success: Original METR HTML saved to .cache/sources/completeness/metr-gpt56-sol.html; SHA-256 recorded; source lines 41–45 checked.