Hackernews posts about Qwen3

RSS feed for Qwen3

31
newRun Qwen 3.8 Flash Next (125B) on consumer hardware (RTX 4090) at 100T/sgithub.com
▲ 923snehesht 2d 💬 421
32
newKev: Tiny Jev-like family of decision models built on top of Qwen3.5github.com
▲ 463tosh 16d 💬 210
33
newQwen 3.8 Omni Flashqwen.ai
▲ 346jjcm 19d 💬 138
34
newBenchmarking Qwen3.8 27B quantizations: 4-bit holds up, 1-bit collapsesquesma.com
▲ 286stared 28d 💬 139
35
newQwen 3.8 follows GPT-5.5 Pro reasoning prefillsgist.github.com
▲ 237wsxiaoys 27d 💬 93
36
newShapelearn Qwen 3.8 27B (13.1 GB VRAM)byteshape.com
▲ 104syntaxing 19d 💬 39
37
newShow HN: Nari Qwen3-TTS and Qwen3-ASR – High accuracy, low latency and costnarilabs.com
▲ 92toebee 22d 💬 31
38
newI made my iPhone a second GPU for my MacBook-Qwen 3.8 27B prefills 29–44% fasterreddit.com
▲ 39thunderbong 3d 💬 2
39
newShow HN: Swift-Qwen3.8-27B, -58.3% thinking, x1.95 speed, accuracy of xhighhuggingface.co
▲ 33kisjovan 20d 💬 17
40
newRun QWEN3.8 27B on 16gb Nvidia GPUsgithub.com
▲ 28Pragmata 19d 💬 10
41
newQwen3.8-Flash-Next non-uniform quantization runs on 2 RTX3090shuggingface.co
▲ 9pfeifferj 22d 💬 2
42
newGVS5H: Five Qwen3.8-27B Models Match Claude Fable 5 on LiveCodeBench Hardgithub.com
▲ 6OakNinja 24d discuss
43
newQwen3.8 models outperforms Fable 5 Using GVS5H Harnessgithub.com
▲ 5pggues 20d 💬 1
44
newTuning a Local Coding Agent: Oh My Pi and Qwen3.8-27B on Two RTX 3090sdoug.sh
▲ 5dougcalobrisi 21d 💬 1
45
newPeer-to-peer LLM inference in browser tabs, Qwen 3.8 27Bgithub.com
▲ 5Nehanth 29d 💬 1
46
newStrata: Qwen3.8-Flash-Next (125B Moe) on a 8GB+ Nvidia GPUgithub.com
▲ 5simonpure 8d discuss
47
newAsk HN: Which harness do you use with Qwen3.8 27B? And why?
▲ 4grigio 13d 💬 10
48
newUnattended 24hr Full-Stack Build: Qwen 3.8 27B on a Single RTX 5090github.com
▲ 4anglepoiselife 15d 💬 1
49
newMining Qwen 3.8 reasoning trace for prompt/skill evaluationolegivye.com
▲ 4olekrk 25d discuss
50
newQwen3.8-Flash-Next Is on TensorFold with Speed Boostsx.com
▲ 3soltanov 7d 💬 2
51
newThinkingCap-Qwen3.8-27B: Less Thinking, Similar Accuracykaitchup.substack.com
▲ 3theanonymousone 10d 💬 1
52
newHalogen – Fastest Way to Run Qwen3.8-Flash-Next on AMD Strix Halogithub.com
▲ 3rdslw 6d discuss
53
newSplash Engine – The Fastest Local Qwen3.8 on Apple Siliconlmstudio.ai
▲ 3sebiw 17d discuss
54
newServe Qwen3 2.4T-Parameter Model with Configurable Reasoning on Nvidia GB300developer.nvidia.com
▲ 3backrun 26d discuss
55
newJev-serve: Run latest qwen3.8, all LLM models, mlx-supportedgithub.com
▲ 2rreinold2 13d 💬 2
56
newPewDiePie uncensors Qwen3.5-9B with Hereticwww.tomshardware.com
▲ 2alanwreath 4d 💬 1
57
newJeff-Code: a 0.8B model makes Qwen 3.8-27B coding 47% faster, same pass rategithub.com
▲ 2firelex 37h discuss
58
newFull-fabric VHDL LLM inference engine. Runs Qwen3.5-class transformer inferencegithub.com
▲ 2jacquesm 2d discuss
59
newGufo-Qwen3.6-35B-A3B-Q6dense – 3095tok/s prefill; 190 tok/s decode on Strix Halogithub.com
▲ 2nubela 4d discuss
60
newShow HN: Moe Routing Atlas – which experts fire in Qwen3.5/3.6 35B-A3Bgslaller.github.io
▲ 2gslaller 7d discuss