SMRTR ProgrammingAug 16, 2025Wired

Developers Say GPT-5 Is a Mixed Bag

SMRTR summary

GPT-5 has received mixed reviews from developers since its launch last week. While it excels at technical reasoning and planning, some developers believe Anthropic's models still produce better code, with GPT-5 sometimes generating unnecessary or redundant lines. The model stands out for being significantly more cost-effective than competitors—costing $30 for tests that would cost $400 with Anthropic's Opus 4.1—though Princeton researchers found it scored lower on accuracy (27% versus Claude's 51%).

SMRTR provides this summary for quick context. The original article belongs to Wired.

Read the original article
SMRTR Programming

Get the next batch of curated summaries in your inbox.

This archive is built from SMRTR newsletter summaries. Subscribe for hand-picked stories without the extra noise.