GPT-6 Astra's SWE results show only ~6% gain over its predecessor, critic notes

robleclerc · x · 2026-09-04

Reacting to GPT-6 Astra's benchmark reveal, robleclerc says he hoped for a much bigger leap in software engineering: the model improved only about 6% over its predecessor on SWE, which he found underwhelming — a sober counterpoint to the launch hype.

Related event: GPT-6 Astra Benchmarks Land with Only ~6% Coding Gain(4 posts)→

Original post →

More from Models

Models channel →