framework
Inference and serving framework listed as compatible with Qwen3.8-27B.
No evidence-backed relationships are recorded.
Moonshot published 2.8 trillion open weights. At four bits per parameter that is about 1.4TB resident before any cache, which rules out the eight-way H100 node most teams assume.
Reality
NVIDIA says Alibaba's largest open-weight model serves over 4K tokens/sec/GPU and 350 tokens/sec/user in FP8 on a GB300 NVL72. That figure is the self-hosting floor, not a benchmark.
Reality
The update adds a path selector and a two-tap convolution rather than layers, recovering most of the accuracy that tripling the drafter bought at 15.2% latency, by the vendor's own numbers.
Reality
A Kubernetes serving guide makes a point most teams skip: the YAML you reviewed is identical across five engines, and everything it hides is what breaks in production.
Reality
A week-long failure log on two RTX 3090s under WSL2 lands on one config at 170-210 tok/s. Everything before it died in dependency resolution, not in the math.
Reality
Alibaba scheduled a 27-billion-parameter vision-language model for August 14, alongside an already-published 2.4-trillion-parameter MoE. The smaller file is the consequential one.
Reality