Reaction
Qwen3.8-Flash-Next is claimed to run at 30-plus tokens per second on devices from all major manufacturers
A user calls the model a sweet spot for capable local devices and links a report on local performance.
TLDR
A user links a report about what can run locally, on which hardware, and how fast. They single out Qwen3.8-Flash-Next, claiming that all major manufacturers have devices capable of running it at 30-plus tokens per second.
Combined views
254
2 Sources, first seen ago
2 likes1 comments4 reposts
