Hackernews posts about GPT

GPT is a type of artificial intelligence language model developed by Meta AI and OpenAI that can generate human-like text based on input prompts.

1. GPT-5-Codex (openai.com)
   391 points by meetpateltech 2 days ago | 134 comments
2. GPT-5 Thinking in ChatGPT (a.k.a. Research Goblin) is good at search (simonwillison.net)
   360 points by simonw 11 days ago | 255 comments
3. Claim: GPT-5-pro can prove new interesting mathematics (twitter.com)
   256 points by marcuschong 28 days ago | 151 comments
4. Addendum to GPT-5 system card: GPT-5-Codex (openai.com)
   249 points by wertyk 2 days ago | 140 comments
5. Tau² benchmark: How a prompt rewrite boosted GPT-5-mini by 22% (quesma.com)
   172 points by blndrt about 16 hours ago | 51 comments
6. From GPT-4 to GPT-5: Measuring progress through MedHELM [pdf] (www.fertrevino.com)
   127 points by fertrevino 27 days ago | 96 comments
7. GPT‑5-Codex and upgrades to Codex (simonwillison.net)
   56 points by amrrs 2 days ago | 15 comments
8. GPT-5 Thinking in ChatGPT (a.k.a. Research Goblin) is shockingly good at search (simonw.substack.com)
   52 points by klelatti 11 days ago | 13 comments
9. GPT-realtime and Realtime API updates (openai.com)
   31 points by meetpateltech 20 days ago | 14 comments
10. Show HN: Spoon-Bending – a framework for analyzing GPT-5 alignment behavior (github.com)
    22 points by pablo-chacon 24 days ago | 3 comments
11. ChatGPT is NOT a LLM – GPT is (www.vincirufus.com)
    18 points by vincirufus 11 days ago | 23 comments
12. Show HN: BrowserOS -- browser agents with GPT-OSS, local llms (github.com)
    12 points by felarof 26 days ago | 2 comments
13. Prompt engineering is collapsing – GPT-5 just proved it
    10 points by yuer2025 24 days ago | 8 comments
14. Show HN: I fine-tuned GPT4.1 on my iMessage history (jonyork.net)
    10 points by jonpizza 21 days ago | 6 comments
15. GPT-5 Thinking in ChatGPT (a.k.a. Research Goblin) is shockingly good at search (simonwillison.net)
    9 points by thunderbong 11 days ago | 2 comments
16. GPT-5 and SQL code generation (www.beekeeperstudio.io)
    7 points by mprev 28 days ago | 1 comments
17. Satya Nadella's Top GPT-5 Prompts (twitter.com)
    6 points by pykello 20 days ago | 9 comments
18. Ask HN: Is GPT-5 a regression, or is it just me?
    6 points by technocratius 4 days ago | 7 comments
19. I treated GPT as a prophet. It answered with a 740-page trilogy (www.scribd.com)
    6 points by gptprophet 25 days ago | 4 comments
20. Show HN: Run gpt-oss-20b on 8GB GPUs (github.com)
    6 points by anuarsh 14 days ago | discuss
21. GPT-5 Is a Terrible Storyteller – and That's an AI Safety Problem (www.christoph-heilig.de)
    6 points by raphman 21 days ago | discuss
22. Study: GPT-4V can evaluate social situations similar to humans (www.utu.fi)
    5 points by giuliomagnifico 8 days ago | 1 comments
23. Mathematical research with GPT-5: a Malliavin-Stein experiment (arxiv.org)
    5 points by FergusArgyll 6 days ago | discuss
24. Show HN: Rebuilding GPT2 inference in ~500 lines of (commented) code (khamidou.com)
    5 points by vorador 23 days ago | discuss
25. GPT-5 Pro is suited for solving hard problems (twitter.com)
    4 points by vismit2000 10 days ago | 7 comments
26. Ask HN: Alternatives or cheaper way to access GPT-5 thinking?
    4 points by zephyruslives 18 days ago | 3 comments
27. Show HN: I built research agents because GPT cant accurately hilite txt in files (app.ubik.studio)
    4 points by ieuanking 8 days ago | 2 comments
28. Nvidia's Nemotron 9B, Multiple Instances, Rivals GPT-5 Pro Performance (twitter.com)
    4 points by mromanuk 8 days ago | 1 comments
29. My 14 year old brother discovered this interesting and repeatable GPT failure (xcancel.com)
    4 points by Alifatisk 12 days ago | 1 comments
30. Long context GPT-OSS fine-tuning (unsloth.ai)
    4 points by danielhanchen 20 days ago | 1 comments