Opus 5.5 reviews other models' code: deletes a lot without breaking anything
CtrlAltDwayne · x · 2026-10-25
Replying to the take that Opus 5.5 isn't much better than Sol 6.1 beyond UI, the author shares a different experience: asking Opus 5.5 to review code written by Sol 6.1 or GPT-family models leads to heavy deletion that barely breaks anything, and they routinely use Opus to clean up the mess GPT models leave behind.
More from coding & agent
- 80,000-agent stress test: async tools push results to queue, ~100KB per idle agent on one EPYC thread — AIFlow_ML · 2026-10-25
- 140 lawyers taught to build their own AI agents in 3 hours, from prompting to personalized analyst briefs — jkubicki · 2026-10-25
- UX designer tests 6 AI website auditors for $5 — same site scores 34, 59 and 73 — noya1979 · 2026-10-25
- AWS open-sources Infinity: a runtime that sleeps idle agents for extreme concurrency — AIFlow_ML · 2026-10-25
- SSH has overstayed its welcome: a dev's wishlist for QUIC, graphics protocol and duplex clipboard — seatedro · 2026-10-25
- Frozen Qwen2.5-72B shows SSD-persistent neural memory with 80/80 bit-exact cross-session reload — Nearby_Indication474 · 2026-10-25