Hackernews posts about Qwen3 30B A3B

1. 01
   

   newQwen3 30B A3B Hits 13 token/s on 4xRaspberry Pi 5github.com

   ▲ 347b4rtazz426d💬 161savesaved
2. 02
   

   newQwen3 30B-A3Bhuggingface.co

   ▲ 87tosh464d💬 27savesaved
3. 03
   

   newQwen3-Coder-30B-A3B-Instructhuggingface.co

   ▲ 32swesnow463d💬 10savesaved
4. 04
   

   newQwen3-30B-A3Btwitter.com

   ▲ 12tosh465d💬 2savesaved
5. 05
   

   newQwen3-VL-30B-A3B-Instruct and Thinkinghuggingface.co

   ▲ 6BUFU399ddiscusssavesaved
6. 06
   

   newQwen3-30B-A3B-2507 and Qwen3-235B-A22B-2507 now support up to 1M token contexttwitter.com

   ▲ 5tosh455ddiscusssavesaved
7. 07
   

   newQwen3-30B-A3B-Instruct-2507huggingface.co

   ▲ 5danielhanchen465ddiscusssavesaved
8. 08
   

   newRunning Qwen3-30B-A3B on ARM CPU of Single-Board Computerold.reddit.com

   ▲ 4tosh556ddiscusssavesaved
9. 09
   

   newLocal AI agent in 44 lines of code powered by Qwen3 30B A3B Instructrentry.co

   ▲ 2ai-christianson448ddiscusssavesaved
10. 10
    

    newSwitching to Qwen3-30B-A3B-Instruct-2507 Significantly Improves Local Agenthexeditreality.com

    ▲ 2llllm463ddiscusssavesaved
11. 11
    

    newQwen3-30B-A3B-Thinking-2507 has been releasedhuggingface.co

    ▲ 2waltercool463ddiscusssavesaved
12. 12
    

    newQwen3-30B-A3B-Thinking-2507twitter.com

    ▲ 2tosh464ddiscusssavesaved
13. 13
    

    newOpencode with Nemotron-3-Nano vs. Qwen3-Coder-30B-A3B vs. GPT-OSS-20B-mxfp4 [video]www.youtube.com

    ▲ 1grigio324ddiscusssavesaved
14. 14
    

    newQwen3-30B-A3B-Thinking-2507 Facehuggingface.co

    ▲ 1tosh464ddiscusssavesaved
15. 15
    

    newAsk HN: HotPin – lossless 120B MoE inference on 24GB RAM (CPU, 50 loc)

    ▲ 14LozzKappa104ddiscusssavesaved
16. 16
    

    newShow HN: Shoehorn, a library to quantize an LLM to fit your Mac's VRAMgithub.com

    ▲ 6rhgraysonii84ddiscusssavesaved
17. 17
    

    newShow HN: Open-Source Browser Use LLM (30B, A3B)huggingface.co

    ▲ 2gregpr07322ddiscusssavesaved
18. 18
    

    newA 30B Qwen model walks into a Raspberry Pi and runs in real timebyteshape.com

    ▲ 363dataminer304d💬 135savesaved
19. 19
    

    newUnsloth improvements to gguf tool calling for Qwen3huggingface.co

    ▲ 1dhorthy463ddiscusssavesaved
20. 20
    

    newShow HN: I run 30B 22tok/s, 109tok/s not novel,6GB/16GB RAM overcoming llama.cppgithub.com

    ▲ 5federicoTXTS101ddiscusssavesaved
21. 21
    

    newLaunch HN: Magnitude (YC S25) – Self-optimizing inference engine for agentsgithub.com

    ▲ 194anerli37d💬 96savesaved
22. 22
    

    newI've tested some local LLMs on prosumer hardware, here are some findings

    ▲ 7felineflock75ddiscusssavesaved
23. 23
    

    newShow HN: Moe-Direct – MoE Models far larger than your RAM, on a consumer desktopgithub.com

    ▲ 1tmxkzm1925-max69ddiscusssavesaved
24. 24
    

    newShow HN: Levi – run AlphaEvolve on your Claude Code/Codex for dirt cheapttanv.github.io

    ▲ 5ttanv151ddiscusssavesaved
25. 25
    

    newShow HN: Nibia Fabric – Turn everyday computers into a shared AI compute clustergithub.com

    ▲ 1abelop39d💬 3savesaved
26. 26
    

    newShow HN: Dicta.to – Local voice dictation for Mac with on-device AIdicta.to

    ▲ 2alamparelli255ddiscusssavesaved
27. 27
    

    newShow HN: Diktafon – Voice memos on cassette tapes, transcribed on-devicegithub.com

    ▲ 1jaromiru107d💬 1savesaved
28. 28
    

    newShow HN: Learn from 30 historical figures, open source, nonprofit, self-hostedgithub.com

    ▲ 49micstradev150d💬 24savesaved
29. 29
    

    newShow HN: Off Grid: On-device AI-web browsing, tools vision,image,voice–3x faster

    ▲ 12ali_chherawalla255d💬 5savesaved
30. 30
    

    newShow HN: I am running 3 coding agents non-stop over the last 3 days. Here is how

    ▲ 10sermakarevich146d💬 3savesaved

· edit