Unverified50% confidenceBenchmarkExact time
The same Gemma 4 E2B model varies from 34.9 to 131.9 tokens/s in single-request generation speed across different AWS inference backends, a significant gap.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/12/2026
First Seen
Valid until: 1/10/2027
Sources
Related Entities
Related Claims
Unverified在本地推理速度对比中，Agent A1（95 tokens/秒）快于Gemma 4 Coder和Qwen系列70% similarUnverifiedThe Laya decision model responds up to 4.1 times faster for short requests, with a 3-question guardrail check on a B200 dropping from 11.6 ms to 2.8 ms.68% similarUnverifiedIn a model comparison test, qwen2.5:3b correctly understood pace, produced 0 hallucinated place names, had a length error of 0.3%, peak memory of 259 MB, and took 106 seconds.67% similarUnverifiedIn Strata's speed table, under a 128k context, Q2_0 (2-bit) achieves 65.1 tokens/s and IQ2_XS achieves 52 tokens/s; under a 4k short context, IQ3_XXS achieves 65.6 tokens/s.66% similarUnverifiedGPTQ（2022年）和AWQ（2023年）能将70B参数的LLM量化到INT4精度，困惑度损失不到1%65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011095API
curl https://kongchang.com/api/v1/knowledge/claims/1011095MCP
get_claim(id=1011095)