4+ day, 17+ hour ago   (138+ words) Everything below is vendor-reported — Tencent's numbers from the UI-Mate-27B card and paper, inclusionAI's from the LLaDA-UI technical report — and none of it has been reproduced by anyone outside the two labs as of September 9, 2026. Treat every figure as a claim....

4+ day, 21+ hour ago   (844+ words) AuK's interface is a single prompt box: you hand the model a natural-language instruction and, where a task needs one, an audio reference, and it returns audio. Grouped by the card's own taxonomy, the five task families are: • Speech generation…...

4+ day, 21+ hour ago   (372+ words) Because the two models do not overlap on any benchmark, the scoreboard is a portrait of two different bets rather than a ranking. Sourcing matters on every row — AuK-Flash's claims are Tencent card statements days old and unreproduced; MathForm-8B's…...

4+ day, 21+ hour ago   (705+ words) • The weights live on Hugging Face under the tencent org and are mirrored on ModelScope under Tencent-Hunyuan — two mirrors, same MIT licence. • Each model repo holds a single safetensors checkpoint plus a config and a VAE: auk_base.safetensors at 6.12 GB and…...

6+ day, 10+ hour ago   (380+ words) The 2B family went public as a rolling, unannounced drop, newest entries first: • 2026-09-06 — the flagship MiniCPM5-2B repo, then MiniCPM5-2B-DSpark twenty-one minutes later. • 2026-09-07 — MiniCPM5-2B-GPTQ, the quantized serving format. The draft model's card reports one figure that matters — acceptance length, the average number of…...

4+ week, 1+ day ago   (558+ words) Unsloth is four separate things, and the keyword search results mix them up. Knowing which one you need is half the setup: • unsloth/Qwen3.8-27B-GGUF — the quantized weight files on Hugging Face, 21 quants plus a vision projector. Built with Dynamic V3.0 (preview),…...