Sonnet 5 aligned an early Opus 4.8 checkpoint

A weaker model can run the alignment training of a stronger one, at least on the failures anyone can measure. Sonnet 5 post-trained an early checkpoint of Opus 4.8 and reached safety scores approaching those of production Opus 4.8, which went through Anthropic’s full alignment training.
In the underlying Fellows project, Claude got 48 hours and 1 GPU to improve the alignment of small models. It researched and proposed methods, then trained and tested the models on its own.
Anthropic says Claude can reliably fix measurable misalignment, though subtle or rare failures may have no benchmark at all. The team is releasing its automated alignment research setup for others to build on.
Up next
PR review backlog pushed Harness to rebuild its Git repo
Some teams are pushing 10x their old code volume, a few claim 50x, and Harness field CTO Martin Reynolds says the pull requests just pile up in review. Harness rebuilt its Git repo for agent traffic and added an AI reviewer.
Related stories
- Anthropic paper: automated researchers fix 10/10 benchmarks
- Claude learned proteins, not just papers about them
- Claude Fable 5.2 hits 69% on PCB layout test, up from 25%
- METR clocks Anthropic benchmark gains at 1.9x old pace
- Claude Fable 5.1 cracks Urquhart’s Cyphral Distich
- The last of Wiedijk’s 100 theorems falls to Anthropic
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

