Hackernews posts about Opus

Opus is an open-source audio codec that has been upgraded with machine learning technology to improve its compression and quality.

  1. GLM 5.2 vs. Opus (techstackups.com)
    519 points by ritzaco 6 days ago | 343 comments
  2. Local Qwen isn't a worse Opus, it's a different tool (blog.alexellis.io)
    492 points by alphabettsy 10 days ago | 253 comments
  3. VibeThinker: 3B param model that beats Opus 4.5 on reasoning with novel SFT+GRPO (arxiv.org)
    396 points by timhigins 5 days ago | 205 comments
  4. Show HN: Formally verified polygon intersection – Opus 4.8 oneshots, prev failed (github.com)
    93 points by permute 23 days ago | 21 comments
  5. Claude: Elevated Error Rates for Opus 4.8, Opus 4.7, Opus 4.6, and Sonnet 4.6 (status.claude.com)
    34 points by forks 6 days ago | 38 comments
  6. Claude Opus 4.8 Max responding to an empty message (xcancel.com)
    27 points by frb 25 days ago | 4 comments
  7. Claude Opus 4.8 distilled Alibaba Qwen models (twitter.com)
    23 points by simjnd 30 days ago | 7 comments
  8. Show HN: OpenHack – OSS security scanner, 40x cheaper, on par with Opus 4.6 (github.com)
    12 points by ananayarora 23 days ago | discuss
  9. Claude Opus 4.8 may have distilled Qwen (old.reddit.com)
    9 points by GaryBluto 29 days ago | 4 comments
  10. Claude Code Degraded Before Opus 4.8 Release (marginlab.ai)
    8 points by qwesr123 30 days ago | discuss
  11. BrokenClaw Part 7: Opus-4.8 Edition – All Emails Lead to RCE (veganmosfet.codeberg.page)
    7 points by e12e 1 day ago | 1 comments
  12. GLM 5.2 vs. Opus (techstackups.com)
    7 points by sixhobbits 9 days ago | 1 comments
  13. Researcher uses Opus 4.8 to find critical counterfeiting vulnerability in Zcash (twitter.com)
    7 points by bryantwolf 23 days ago | 1 comments
  14. Stats from 30K AI debates: Opus 4.7 is the most influential model (opper.ai)
    7 points by felix089 24 days ago | 1 comments
  15. GLM-5.2 vs. Claude Opus 4.8: Full Comparison (llm-stats.com)
    7 points by gjvc 8 days ago | discuss
  16. Elevated error rate on Claude Opus 4.8 (status.claude.com)
    6 points by lvncelot 4 days ago | 2 comments
  17. Opus 4.8 feels worse then sonnet
    5 points by mrprincerawat 3 days ago | 5 comments
  18. GLM 5.2 vs. Claude Opus 4.5 (gopeekapp.blogspot.com)
    5 points by GeorgeWoff25 5 days ago | 2 comments
  19. Chrome, Opera and Vivaldi pen open letter to Microsoft saying "enough is enough" (www.neowin.net)
    5 points by bundie 23 days ago | 1 comments
  20. DeepSWE: More and cheaper intelligence from maxed GPT 5.5 than maxed Opus 4.8 (twitter.com)
    5 points by rajveerb 28 days ago | 1 comments
  21. Show HN: LMAO – Temu League of Legends, Built with Opus 4.8 (lmaomoba.com)
    4 points by jonnyasmar 25 days ago | 4 comments
  22. Stats from 30K AI debates: Opus 4.7 is the most influential model (opper.ai)
    4 points by felix089 23 days ago | 3 comments
  23. Local Qwen isn't a worse Opus, it's a different tool (blog.alexellis.io)
    4 points by alexellisuk 11 days ago | 2 comments
  24. Claude Opus is more performant on OpenCode than Claude Code (artificialanalysis.ai)
    4 points by log101 17 days ago | 1 comments
  25. GLM-5.2 vs. Claude Opus: Same Code, Less Than Half the Cost (entelligence.ai)
    4 points by Entelligence25 4 days ago | discuss
  26. Show HN: Claude Opus 4.8 Masterclass – Effort control and dynamic workflows (ddsboston.com)
    4 points by robert_dds 26 days ago | discuss
  27. Same prompt. Different teammate. My 5 cents on Opus 4.8 (norahsakal.com)
    4 points by norsak 28 days ago | discuss
  28. Was GLM-5.2 trained on Opus 4.5 outputs? (1chat.com)
    3 points by coolspot 1 day ago | 3 comments
  29. Ask HN: Did Anthropic Nerf Opus 4.8?
    3 points by Ancalagon 17 days ago | 2 comments
  30. Local AI model claim to beat GPT 5.5 and Opus 4.7 (old.reddit.com)
    3 points by trilogic 20 days ago | 2 comments
More