Unverified50% confidenceBenchmarkExact time
The 12-billion-parameter Gemma dense model achieved a measured speed of just over 10 tokens/second.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/19/2026
First Seen
Valid until: 12/18/2026
Sources
Related Entities
Related Claims
Unverified在11.8万节点的生产级图基准测试中，Semantica节点搜索耗时从24毫秒降低到0.004毫秒，约6000倍加速67% similarUnverified投机解码的理想草稿模型倍速区间应在10到30倍之间，通常意味着草稿模型参数量约为目标模型的1/50至1/1065% similarUnverifiedIn actual testing, a single classification by the Dify question classifier takes approximately 11 seconds, because execution speed depends on the large model's response.64% similarUnverifiedGemma's token generation speed on high-performance laptops can reach 20–50 tokens per second.63% similarUnverifiedPilot experiments had three coding agents and eight models each run eight 20-minute single-pass tasks, with label-aware F1 showing extremely high variance ranging from 0.143 to 0.892.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/933719API
curl https://kongchang.com/api/v1/knowledge/claims/933719MCP
get_claim(id=933719)