Unverified50% confidenceDecision RuleExact time
For scenarios that pursue generation throughput and are tight on VRAM, NVFP4 decodes 2.3 times faster on the Ada architecture with weights of only 3.28 GiB, but one must accept 89%–90% token-level fidelity.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/18/2026
First Seen
Valid until: 1/16/2027
Sources
Related Entities
Related Claims
UnverifiedIn this test, NVFP4 achieved a generation speed of 116.8 tok/s, 2.3 times that of BF16 (51.0 tok/s), and the NVFP4 weight size was 3.28 GiB.74% similarPartially VerifiedNVFP4 occupies 4 bits per parameter, i.e., 0.5 bytes, theoretically halving weight memory usage and load bandwidth compared with FP8.70% similarUnverifiedNVFP4 is actually approximately 3.75 bits, and is not strictly a 4-bit quantization in the rigorous sense.68% similarUnverifiedThe RTX 3090 cannot run the NVFP4 quantization format because its Ampere architecture Tensor Cores do not support NVFP4 instructions, leaving only the Q4 format available, with speeds of 40–48 tokens/second.65% similarUnverifiedOn NVIDIA L4, the bf16 model (6.32 GiB) has a theoretical memory-bandwidth ceiling of about 47 tokens/s, while the int4 model (2.85 GiB) has a theoretical ceiling of about 105 tokens/s, with measured peak reaching 131.9 tokens/s.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1057663API
curl https://kongchang.com/api/v1/knowledge/claims/1057663MCP
get_claim(id=1057663)