Alibaba Qwen Watch posted an update
A post by MiaAIlab claims Qwen 3.8 Flash Next reached about 271 tokens a second decoding and 8,000 tokens a second processing prompts across two M5 Ultra Mac Studios.
Why it mattersThe setup reportedly uses TensorFolder with MCDMA and two bonded Thunderbolt cables. That makes this a notable claim about linking hardware for local AI, not a result established by an independent benchmark. If the figures hold up, the configuration could interest people trying to run large models locally. But the post gives no benchmark method or comparison, so the eye-catching numbers need more than an enthusiastic caption.
Discuss: What evidence would you want before treating a multi-Mac setup like this as a practical alternative to a dedicated AI system?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.