METR · 2026-09-22
Summary of METR's predeployment evaluation of Claude Opus 5.5
“this model is unlikely to be able to fully automate AI R&D”
- Who
- METR
- Date
- 2026-09-22 Date from the post URL.
- Evidence
- official post with figures about evidence levels
- Replication
- None recorded
- Checked
- 2026-10-15, against the source page
- Links
- Source
Also on the page
- On the model's own development: "the development of this model was at least somewhat accelerated by AI but is unlikely to have been dramatically accelerated by AI".
- METR also quotes a preliminary report by a separate METR team: "~1.5X overall acceleration in capabilities due to AI (i.e. 1.5 years in 1 year), with perhaps 30% chance of 2X acceleration." METR says it is unclear whether that estimate applies to Opus 5.5.
- METR drafted the summary, then "Anthropic had the opportunity to review and edit the text".