Hackernews posts about DeepSeek v3

  1. DeepSeek-V3 Technical Report (arxiv.org)
    132 points by signa11 26 days ago | 34 comments
  2. DeepSeek releases their latest DeepSeek v3 model, now featuring an MIT license (simonwillison.net)
    34 points by spenvo 29 days ago | 1 comments
  3. Deepseek V3-0324 (huggingface.co)
    18 points by bananaflag 29 days ago | 2 comments
  4. DeepSeek-V3-0324 released, 641GB, MIT licensed, >20tok/SEC on $<10k hardware (web.archive.org)
    15 points by Thoreandan 28 days ago | 4 comments
  5. DeepSeek V3 is now the highest scoring non-reasoning model (twitter.com)
    14 points by aurareturn 28 days ago | 3 comments
  6. DeepSeek-V3-0324 Crushes GPT-4.5 in Math and Code Benchmarks at 1/277 the Cost (api-docs.deepseek.com)
    10 points by kong75 28 days ago | 2 comments
  7. DeepSeek-V3 now runs at 20 tokens per second on Mac Studio (venturebeat.com)
    9 points by amrrs 29 days ago | discuss
  8. DeepSeek-V3 runs at 20 tokens/s on Mac Studio, and that's a nightmare for OpenAI (venturebeat.com)
    8 points by benocodes 28 days ago | 2 comments
  9. Show HN: Trickle – A coding agent for non-devs, now with free DeepSeek V3 (www.trickle.so)
    7 points by jarodxuming 28 days ago | 2 comments
  10. DeepSeek V3 0324 outpaces GPT 4.5 and Claude 3.7 in coding, other benchmarks (huggingface.co)
    7 points by xianshou 27 days ago | discuss
  11. DeepSeek v3: Advanced AI Language Model (deepseekv3.org)
    5 points by vismit2000 28 days ago | 1 comments
  12. DeepSeek-V3-0324 (huggingface.co)
    5 points by desideratum 29 days ago | 1 comments
  13. New DeepSeek V3 0324 with MIT license (huggingface.co)
    5 points by AbuAssar 28 days ago | discuss
  14. DeepSeek-V3 now runs on Mac Studio, and that's a nightmare for OpenAI (venturebeat.com)
    4 points by tu7001 28 days ago | discuss
  15. How to Run Deepseek-V3-0324 Locally (unsloth.ai)
    3 points by tosh 24 days ago | discuss
  16. Run DeepSeek-V3-0324 locally (www.unsloth.ai)
    3 points by tomdekan 27 days ago | discuss
  17. DeepSeek V3-0324 Posted to HuggingFace (huggingface.co)
    3 points by jlaneve 29 days ago | discuss
  18. DeepSeek V3-0324: New DeepSeek model released (medium.com)
    2 points by wertyk 29 days ago | discuss
  19. DeepSeek v3 update coming to Hyperbolic
    2 points by dubrado 29 days ago | discuss
  20. Thread comparing M4 Mac Studio 512GB vs. PC build running Deepseek-v3 671B-Q8 (old.reddit.com)
    1 points by bredren 21 days ago | 1 comments
  21. 2.71bit DeepSeek-V3-0324 (unsloth.ai)
    1 points by danielhanchen 27 days ago | 1 comments
  22. DeepSeek V3 0324 is now the best nonthinking model (Reddit) (old.reddit.com)
    1 points by xianshou 27 days ago | discuss
  23. DeepSeek V3 (non-thinking) benchmark (old.reddit.com)
    1 points by aurareturn 28 days ago | discuss
  24. Show HN: Qwen-2.5-32B is now the best open source OCR model (github.com)
    211 points by themanmaran 21 days ago | 47 comments
  25. Show HN: QwQ AI – Aggregator of free LLMs answer generator (qwq32.com)
    6 points by zhangchengzc 3 days ago | 1 comments
  26. Show HN: Plandex v2 – OSS AI coding agent w/ diff sandbox, full auto, 2M context (plandex.ai)
    4 points by danenania 19 days ago | discuss
  27. Show HN: Tokiwi – An online tokenizer for any Hugging Face model (www.tokiwi.dev)
    2 points by TweedBeetle 24 days ago | discuss
  28. DeepSeek releases new V3 checkpoint (V3-0324) (old.reddit.com)
    3 points by nsoonhui 28 days ago | discuss
  29. Show HN: ʻŌlelo Honua v3.1.2 – Automating internationalization (www.olelohonua.com)
    3 points by rbrownmh 22 days ago | discuss
  30. DeepSeek-V3 (github.com)
    125 points by mgt19937 4 months ago | 40 comments
  31. Notes on the New Deepseek v3 (composio.dev)
    99 points by soham123 4 months ago | 24 comments
  32. DeepSeek v3 beats Claude sonnet 3.5 and way cheaper (huggingface.co)
    48 points by helloericsf 4 months ago | 9 comments
  33. DeepSeek V3 and the cost of frontier AI models (www.interconnects.ai)
    32 points by walterbell 3 months ago | 9 comments
  34. DeepSeek-V3/R1 Inference System Overview (github.com)
    27 points by meetpateltech about 2 months ago | 6 comments
  35. DeepSeek-v3 Technical Report [pdf] (github.com)
    26 points by pongogogo 4 months ago | 2 comments
  36. Scaling the Tülu 3 post-training recipes to surpass the perf of DeepSeek V3 (allenai.org)
    14 points by Philpax 3 months ago | discuss
  37. DeepSeek v3: The Six Million Dollar Model (thezvi.substack.com)
    12 points by paulpauper 4 months ago | discuss
  38. Meta GenAI team is in panic mode due to deepseek v3 (twitter.com)
    11 points by cpt100 3 months ago | 3 comments
  39. Alibaba releases AI model it claims surpasses DeepSeek-V3 (finance.yahoo.com)
    10 points by mgh2 3 months ago | discuss
  40. Running DeepSeek V3 671B on M4 Mac Mini Cluster (blog.exolabs.net)
    8 points by chadash 4 months ago | 3 comments
  41. DeepSeek-V3 (api-docs.deepseek.com)
    8 points by mromanuk 4 months ago | discuss
  42. DeepSeek-V3, ultra-large open-source AI, outperforms Llama and Qwen on launch (venturebeat.com)
    7 points by hochmartinez 4 months ago | discuss
  43. DeepSeek-V3 Chat is out (chat.deepseek.com)
    5 points by sss111 4 months ago | discuss
  44. Ask HN: Deepseek v3 how can I check on local computer?
    4 points by lavren1974 4 months ago | 9 comments
  45. DeepSeek-V3-Base (huggingface.co)
    4 points by pr337h4m 4 months ago | discuss
  46. New Qwen2.5-Max Outperforms DeepSeek V3 in Benchmarks (twitter.com)
    3 points by helloericsf 3 months ago | 2 comments
  47. DeepSeek-V3 (twitter.com)
    3 points by doener 4 months ago | 2 comments
  48. DeepSeek-V3, ultra-large open-source AI, outperforms Llama and Qwen on launch (venturebeat.com)
    3 points by cannibalXxx 4 months ago | 1 comments
  49. Day 6 of OpenSourceWeek: One More Thing–DeepSeek-V3/R1 Inference System Overview (twitter.com)
    3 points by tempaccount420 about 2 months ago | discuss
  50. DeepSeek-V3-Turkish
    3 points by can-deliktas 2 months ago | discuss
  51. DeepSeek-V3 Technical Report (arxiv.org)
    3 points by doener 3 months ago | discuss
  52. Running DeepSeek V3 671B on M4 Mac Mini Cluster (blog.exolabs.net)
    3 points by akyuu 4 months ago | discuss
  53. Why does DeepSeek v3 route to 8 experts per token? (praneet.sh)
    2 points by prrathi 3 months ago | 1 comments
  54. Ask HN: DeepSeek V3's AI Code Review Performance – A Reality Check with Data
    2 points by Jet_Xu 4 months ago | 1 comments
  55. You can run DeepSeek-V3 for 24 hours and it will cost you $1.52 (xcancel.com)
    2 points by Alifatisk 4 months ago | 1 comments
  56. DeepSeek-V3 has a problem: it keeps claiming to be ChatGPT (www.neowin.net)
    2 points by bundie 4 months ago | 1 comments
  57. Why DeepSeek v3 was such a big deal (edforson.substack.com)
    2 points by edshiro about 1 month ago | discuss
  58. Show HN: Showdown of Titans: GPT-4o vs. DeepSeek V3 in AI Debate Arena (v0-ai-debate.vercel.app)
    2 points by Sai_Praneeth 3 months ago | discuss
  59. Alibaba releases AI model it claims surpasses DeepSeek-V3 (ca.finance.yahoo.com)
    2 points by thunderbong 3 months ago | discuss