qwen
23 posts tagged qwen.
- FP4 KV Cache Done Right: What BitDecoding-FP4 Changes for Long-Context Serving
- Tencent’s Qwen DSpine hint is about inference cost, not model size
- Arete-27B tests values as a post-training layer
- WebHarbor Puts 153 Fake Real Websites in a Box for Agents to Train On
- A 0.6B Persian Embedding Model That Catches a 4B: What Yalda Teaches About Distillation
- A 4B Qwen fine-tune points at local 3D scene generation
- When a Benchmark Claim Has No Benchmark Behind It
- WeAct Turns LoRA Adapters Into Something an Agent Can Pick Mid-Task
- What greedy sampling and expert-skip defaults cost you on a single-4090 Qwen3.8 rig
- What SpaceCast-Bench reveals about the gap between seeing space and predicting it
- When the Prompt's Phrasing Changes the Answer: SCAPO and Token-Level Credit in RLVR
- Qwen 27B on a 4090 is a useful reminder about local AI demos
- Qwen architecture rumors do not make a 3090 fast by default
- Audio models do not automatically share speech and text concepts
- The Skill Router You Already Have: Gavel Reads Routing From a Frozen LLM
- MindTopo Tests Whether Vision Models Grasp Topology, Not Just Distances
- Qwen 3.8 Flash Next shows why chat feel still matters
- Qwen 3.8’s benchmark buzz is about reasoning budget, not vibes
- Qwen thinking levels need task-level tests, not vibes
- Apple’s reported Alibaba deal is a China AI reality check
- Qwen3.8-27B looks useful as overnight local coding labor
- Qwen’s BASIC ray-tracer demo is really about closed-loop coding
- Qwen’s next move matters after the weights land