Lovable Adopts Opus 5.5: Same Quality at Up to 48% Fewer Steps
AlexandrePesant · x · 2026-09-23
Lovable announced Opus 5.5 is now in its platform: matching Opus 5 quality on 0-to-1 building and iterative code fixing while using up to 48% fewer steps on existing codebases. Self-verification scores are 4-6% higher with up to 57% fewer steps.
Gains come from one-pass context gathering, fewer but more complete edits, and no unnecessary re-checking. The post details Lovable's internal benchmark: 0-to-1 building, fixing/evolving live codebases, and verification discipline (checking exactly as much as needed), with each task run 3+ times at equivalent reasoning effort levels.
More from coding & agent
- Grok Build 1.0.41 Ships with Grok 4.7, Adds Subagent Controls — elonmusk · 2026-09-23
- Lovable routes traffic to GPT-6 Sol and Opus 5.5, citing capability leap — w_hgm · 2026-09-23
- Every's GPT-6 Sol Hands-On: Near-Astra Writing, 50% Cheaper Than 5.6 Sol — every · 2026-09-23
- Opus 5.5 is now live in Conductor — charlieholtz · 2026-09-23
- Reproducible agent evals: harbor makes configs, trajectories and logs shareable — seanwbren · 2026-09-23
- AntLing open-sources Ming-Image-0.1-Design, tops open-weight UI design leaderboard — niacolhealth · 2026-09-23