← Back to archive

Composer 2.5 tops CursorBench at 20x lower cost than Opus

Oh wow! Composer 2.5 is #3 in CursorBench evals and 20x times cheaper than Opus 4.7 Max.

It's below GPT 5.5 Low, Opus 4.7 Low.

... and Kimi (funny that it's fork of Kimi 2.5)

Data is everything, especially when you’re training on millions of real coding attempts with outcome signals.

Originally posted on X

PS: I'm pretty active on 𝕏 if you'd like to follow along

Archive

2026

2025

2024

2018