deepseek-v4.1POST
vLLM matches DeepSeek's own engine on V4.1 Flash decoding
Claims vLLM now matches DeepSeek's in-house inference engine on V4.1 Flash decoding, hitting 954K tokens/s per GPU on SemiAnalysis AgentX.
artem
@sunset_evmArchived 4 days agoOriginally posted 2026.10.26
artem @sunset_evmOriginally posted 2026.10.26
vLLM now matches DeepSeek's own inference engine on DeepSeek-V4.1-Flash decoding, at 954K tokens/s per GPU on SemiAnalysis AgentX.
This post quotes another post: x.com/vllm_project/status/2114054277379641770
x.com/sunset_evm/status/2114396591314608321Archived 2026.10.26 12:04