Unverified50% confidenceSolutionExact time
The recommended way to run DeepSeek V4.1 Flash on Gorgon Halo is to use the Dwarf Star implementation with a Q2 quantized version, where the main model weights occupy approximately 163GB and the complete file is approximately 366GB (including the Ngram table).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/5/2026
First Seen
Valid until: 1/3/2027
Sources
Related Entities
Related Claims
UnverifiedAfter compression, the DeepSeek V4 Flash model size was reduced from nearly 500 GB to 80.8 GB.71% similarUnverifiedStrix Halo的统一内存架构带宽可达256GB/s70% similarUnverifiedAccording to community-cited information, DeepSeek-V4.1-Flash has 552B backbone parameters, 196B Engram parameters, and safetensors weight files totaling approximately 350 GB.70% similarUnverifiedGorgon Halo retains 16 Zen 5 CPU cores and 40 RDNA 3.5 GPU compute units, with GPU frequency increasing from 2.9GHz to 3GHz and theoretical memory bandwidth increasing from 256GB/s to approximately 273GB/s (approximately 7%).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.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/977947API
curl https://kongchang.com/api/v1/knowledge/claims/977947MCP
get_claim(id=977947)