Unverified50% confidenceBenchmarkExact time
On the Python side, single-prompt model.generate speeds up Llama and Qwen3 models by up to 2.7x at 32K context; with bitsandbytes 4-bit models, generating 2 to 4 prompts at a time speeds up by up to 2.8x on some NVIDIA GPUs.
1
Sources
50%
Confidence
Short-term (~14 days)
Relevance
10/15/2026
First Seen
Valid until: 10/29/2026
Sources
Related Entities
Related Claims
UnverifiedQwen3.5 with approximately 800 million parameters achieved the fastest generation speed, approaching 15 tokens per second, with the second-place LFM2.5 exceeding 10 tokens per second.76% similarUnverifiedUsing Q4 quantized weights (a modified version of llama.cpp), the generation speed on the first run was only 1.83 tokens per second.69% similarUnverifiedThe typical generation speed of an NVIDIA RTX 4090 running a similarly scaled Q4 quantized model is in the range of 40–60 TPS69% similarUnverifiedThe UP host tested MiniCPM-2B on a computer with 4G VRAM and achieved a generation speed of 57 tokens per second, with a first-token latency of only 0.22 seconds.69% similarUnverifiedQwen3 27B's local inference speed is approximately 10 tokens per second.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1039705API
curl https://kongchang.com/api/v1/knowledge/claims/1039705MCP
get_claim(id=1039705)