My benchmark scored GPT-5.4 mini 1.00 by silently skipping the 3 questions it failed

Originally published at cchinchilla.dev. Part 6 of From code to weights, a 12-part series on ML...

Here's a line from YouTube's auto-captions on a lecture about attention in neural...

I ran one prompt through four models last week. Opus 5.5 and Sonnet 5.5 in Claude Code, GPT-6 Astra...

Opus 5.5 pricing: $4/$20 per million tokens, 60% off cache reads. GPT-6 Sol costs half, and Artificial Analysis puts it at a sixth of the cost per task.

OpenAI DevDay 2026 recap: dots, GPT-6.1 Sol at $2/$10, Ultrafast, Pro 500, Decisions API, Codex cloud, Agents API and Sign in with ChatGPT.

Vibe coding at the limit: Armin Ronacher let GPT-6 Astra run alone for 35 hours. It cost about $1,200 and wrote 75,000 lines of nothing of value.

Building a GPT application looks deceptively easy. The first version might be twenty lines of...

Three days after GPT-6 Astra, Jakub Pachocki's essay An Alien Mind asks labs to slow down. OpenAI's own chart shows its safety pause moved GPUs, not cut them.

LibGen in the OpenAI lawsuit: the authors' brief says OpenAI trained GPT-3 on pirated books, told Bill Gates, then deleted them. OpenAI says fair use.

On September 22, 2026, Anthropic shipped Claude Opus 5.5 and OpenAI shipped GPT-6 Sol on the same...