AnalysisAI ModelsJuly 28, 2026
DeepSeek V4 Flash hits 32 tok/s on AMD Ryzen AI MAX+ 395
A user ran DeepSeek V4 Flash plus its speculative draft on a single AMD Ryzen AI MAX+ 395 with 128 GB unified memory, reaching up to 32 tok/s decode. Details in linked blog post.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Record chip earnings fail to satisfy AI investors as stocks slide
- PolyAI releases Dialog-RSN-1 audio-native dialog model
- Build a policy-governed multi-agent financial workflow with Omnigent
- H3 enters arena; MiniMax open weights coming soon
- Thinking Machines releases Inkling-Small multimodal model