Unverified50% confidenceBenchmarkExact time
Developer Soto wrote his own inference engine and ran the 552B-parameter DeepSeek V4.1 Flash on a gaming PC with a single 24G RTX 4090, 90G of memory plus NVMe storage, measuring about 27 tokens/s decoding and 180-320 tokens/s prefill.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/18/2026
First Seen
Valid until: 11/1/2026
Sources
Related Entities
Related Claims
UnverifiedAccording to community-cited information, DeepSeek-V4.1-Flash has 552B backbone parameters, 196B Engram parameters, and safetensors weight files totaling approximately 350 GB.73% similarUnverifiedDeepSeek-V4-Flash-Vision-Exp adopts a MoE architecture with approximately 305B total parameters, activating only approximately 13B per token, for an activation rate of approximately 4.6%.70% similarUnverifiedThe UD-Q8_K_XL quantized version of DeepSeek V4 Flash has a file size of 162 GB, uses MXFP4+BF16 mixed precision, and achieves completely lossless quality.70% similarUnverifiedQwen3-8B Flash performs comparably to DeepSeek's 27B model, has a native context of 256K expandable to 1M, and the official website already provides an API interface with 1 million context.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1059794API
curl https://kongchang.com/api/v1/knowledge/claims/1059794MCP
get_claim(id=1059794)