event
Qwen Update
-
Not my OC. From Universal Modder on X: "Introducing NPCTalk: make the NPCs of any game talk in realtime.
-
Tested 18 new Local LLMs on two RTX 5060 Ti cards: none replaced Qwen3.8-27B (siimtolk.github.io via hn) 3w
I let an AI agent test local LLMs while I sleep An AI agent tested 18 new open-weight models on my home server overnight. None replaced the Qwen3.8-27B my assistant already runs.
-
Long-horizon agentic reinforcement learning (RL) requires both efficient rollout generation and training on long sequences. However, the memory requirements and model-specific operators of Qwen3.8-Flash-Next make this difficult with existi…
-
A new Speculative decoding drafter for vLLM and SGLang called EAGLE-4 (github.com via hn) vllm 3w
EAGLE-4 One Flexible Drafter for All Batch Sizes and Hardware Paper (EAGLE-4) · EAGLE-4 draft for Qwen3.8-27B · Documentation Figure 1(a) of the EAGLE-4 paper: Qwen3.8-27B on the SPEED-Bench qualitative split, 4× NVIDIA B300, one request,…
-
Reposting again since original post got removed by Reddit for some reason (I have no idea why) A few of you pointed out in my last post that comparing Flash-Next and Opus with different harnesses and effort levels wasn't fair. So here is a…
-
An interactive site plotting Artificial Analysis Intelligence vs parameter count (mr-september.github.io via reddit) moe 3w
https://preview.redd.it/mu3u3oqzzrwh1.png?width=2136&format=png&auto=webp&s=ea25af95068ab1c09d5dfe048944fc39ed67f278 Artificial Analysis scores models on its Intelligence Index, but the charts put it against price and speed. For local use…
-
Every prompt sent to a hosted language model is readable by whoever runs it. We study whether the prompt can remain on the client while the host receives transformed vectors that stand in for it.
-
Best LLM Fine-Tuning Services: What to Look for in 2026 (www.rebiha.com via hn) gemmallamafine-tuning 3w
Pick a model — Qwen3.8, Granite, OLMo, Gemma 4, Llama & more — choose from 54 ready-made datasets, and train. No data prep, no infrastructure, no API.
-
2x Tesla p100s, q6_k quant, Qwen 3.8 27B ~60tps (github.com via hn) llamaqwen 4w
This is a fork of llama.cpp tuned for two Tesla P100s (sm_60). On Qwen3.8-27B, plain decode runs at 1.86x upstream's speed (17.5 -> 32.6 t/s), and ~3x with the model's built-in MTP head (~54 t/s).
-
Show HN: Capy-flash – a 125B MoE model on two 16 GB consumer GPUs (github.com via hn) moe 4w
Qwen3.8-Flash-Next (125B MoE) in NVFP4 on two 16 GB consumer GPUs. Against Strata 0.1.44.1 on the same machine (2x RTX 5070 Ti 16 GB, Ryzen 9 9900X, 96 GB DDR5): | | capy-flash, NVFP4 | Strata, same NVFP4 weights | Strata, 3-bit IQ3_S | |-…