Unverified50% confidenceBenchmarkExact time
In 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.
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 a prefill-only test running gpt-oss-120b on 4× B200s, DWDP4 achieves a 1.92× speedup over DEP4 under the MNT 32K/ISL 32K configuration and a 1.54× throughput improvement at saturation (506K vs. 329K tok/s).77% similarUnverifiedFor 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.74% similarUnverifiedNVFP4 is actually approximately 3.75 bits, and is not strictly a 4-bit quantization in the rigorous sense.72% similarUnverifiedThe typical generation speed of an NVIDIA RTX 4090 running a similarly scaled Q4 quantized model is in the range of 40–60 TPS72% similarUnverifiedOn an RTX 3060, the developer's hand-written Triton fused W4A16 GEMV kernel targeting batch-1 decoding runs at about 25.5 tok/s, matching the roughly 25.7 tok/s of the bitsandbytes NF4 quantization scheme.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1057656API
curl https://kongchang.com/api/v1/knowledge/claims/1057656MCP
get_claim(id=1057656)