Unverified50% confidenceFactExact time
The OCP-standard MXFP4 uses E8M0 (pure exponent) scaling factors with a block size of 32.
1
Sources
50%
Confidence
Long-term
Relevance
10/27/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedNVFP4 is a block-scaled 4-bit floating-point format defined by NVIDIA, where every 16 numbers share one FP8 scale; MXFP8 is an 8-bit block floating point defined by OCP, where every 32 numbers share one E8M0 scale.73% similarVerifiedThe MXFP4 format has a group of elements (typically 32) sharing a single 8-bit scaling factor, with each element represented by a 4-bit micro floating-point number (1 sign bit + 2 exponent bits + 1 mantissa bit).73% similarVerifiedMXFP4 mitigates numerical overflow and precision degradation at 4-bit low precision by sharing a single scaling factor within small blocks of data (typically 32 elements), theoretically reducing memory footprint and bandwidth requirements by approximately 4× compared to FP16.70% similarUnverifiedThe 5M model's configuration includes: 21 quantile nodes from 0.01 to 0.99, context length 32768, and patch size 32.63% similarVerifiedFP8 is an 8-bit floating-point format natively supported by newer architectures such as the NVIDIA H100, which can compress data volume by half compared to FP16/BF1663% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1101736API
curl https://kongchang.com/api/v1/knowledge/claims/1101736MCP
get_claim(id=1101736)