Unverified50% confidenceBenchmarkExact time
在预填充阶段Ornith速度接近每秒200个Token,比Qwen快约20个Token
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/9/2026
First Seen
Valid until: 10/7/2026
Sources
16G显存实测:Ornith 35B vs Qwen 35B全方位对决
bilibili程序员-智能译站7/7/2026
Related Claims
UnverifiedQwen 3.6在测试中预填充速度达到400 tokens/秒,生成速度接近40 tokens/秒72% similarUnverified在输出(Decode)速度测试中,2 万 token 以内可达约 45 tokens/秒,3 万 token 以上骤降至约 20 tokens/秒63% similarUnverified启用MTP后推理速度波动显著,有时超过40 tokens/秒,有时降至25 tokens/秒62% similarUnverified本地7B模型的生成速度通常在每秒20-50个token,而云端API可以达到每秒100+token62% similarUnverified通常认为单并发 100 tokens/秒以上即可实现流畅的打字机效果,约 200 tokens/秒生成 2000 字长文本仅需约 10 秒61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/342891API
curl https://kongchang.com/api/v1/knowledge/claims/342891MCP
get_claim(id=342891)