GPT-5 Has Arrived

Why It Matters

Making GPT-5 broadly available means up to a billion users gain access to a materially stronger AI for coding, technical and multimodal tasks, potentially shifting developer and enterprise adoption and pressuring rivals like Anthropic. Its mix of improved utility, lower cost and wide distribution could accelerate AI integration into consumer and professional workflows while keeping limits on hallucinations and context size that temper expectations.

Summary

OpenAI has released GPT-5 to free-tier ChatGPT users, delivering noticeable gains in coding, multimodal reasoning, and reduced hallucinations versus prior models, though it is not a breakthrough AGI. Early tests show strong performance on certain logic and software benchmarks—outperforming competitors like Anthropic on SweetBench verified and beating Gemini DeepThink on some image-and-chart tasks—while overall accuracy on public logic tests hovers around the high 50s percent in the reviewer’s runs. The model still hallucinates roughly 5% on common user queries, offers only modest increases to context window size, and benefits most where its training data is rich. OpenAI emphasized real-world utility, affordability, and mass accessibility over pushing maximal capability at this stage.

Original Description

GPT-5 will change how hundreds of millions of people use AI. Yes, you might have to forgive the chart crimes, the underwhelming livestream and Altman hype… But it’s a good model. I have read the 50 page system card in full, have the benchmark scores, coding tests, and things you might have missed.

https://app.grayswan.ai/ai-explained

AI Insiders ($9!): https://www.patreon.com/AIExplained

Announcement: https://openai.com/index/introducing-gpt-5/

System Card: https://cdn.openai.com/pdf/8124a3ce-ab78-4f06-96eb-49ea29ffb52f/gpt5-system-card-aug7.pdf

Extra Paper: https://cdn.openai.com/pdf/be60c07b-6bc2-4f54-bcee-4141e1d6c69a/gpt-5-safe_completions.pdf

Altman tweet: https://x.com/sama/status/1953551377873117369

Livestream: https://www.youtube.com/watch?v=0Uu_VJeVVfo

METR Report: https://metr.github.io/autonomy-evals-guide/gpt-5-report/

ARC-AGI-2: https://x.com/fchollet/status/1953511631054680085

Claude Opus 4.1: https://www.anthropic.com/news/claude-opus-4-1

MMMU: https://mmmu-benchmark.github.io/

Cursor Praise: https://x.com/ryolu_/status/1953531724895596669

Non-hype Newsletter: https://signaltonoise.beehiiv.com/

Podcast: https://aiexplainedopodcast.buzzsprout.com/