Unverified50% confidenceBenchmarkExact time
MIMO在MXFP4下提示处理起始约477 tokens/s,32K时降至354,token生成从18降到16
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/5/2026
First Seen
Valid until: 10/19/2026
Sources
Related Entities
Related Claims
Unverified在 Qwen3 27B 的 llama.cpp 环境下,R9700 跑出 27.1 tokens/s(达带宽上限约65%),B70 跑出 21.4 tokens/s(约50%出头)64% similarUnverified对照模型Supra-50M有51.8M参数,bf16下103.6MB,8-bit量化后56.2MB,降到int4困惑度从165恶化到193,压到三元直接崩溃62% similarUnverified在4bit量化版本(完全放入32GB显存)测试中,Ollama达240 tokens/s,FreeToken为225 tokens/s,Ollama反而更快62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/977950API
curl https://kongchang.com/api/v1/knowledge/claims/977950MCP
get_claim(id=977950)