Unverified50% confidenceBenchmarkExact time
Dual-card comparison experiments: Pro 4500 running Q4 went from 61.5 tok/s on a single card to 111.0 tok/s on dual cards (1.81x); Tesla P40 went from 19.6 to 34.8 tok/s (1.78x); the other two groups were 1.48x and 1.22x.
1
Sources
50%
Confidence
Long-term
Relevance
10/24/2026
First Seen
Sources
141 Real-World Benchmarks Revealed: How Fast Can Your GPU Really Run Large Models?
bilibili九物AI生产力10/23/2026
Related Entities
Related Claims
UnverifiedAt a 128K context, Q2_0 outputs 65.1 tok/s, IQ2_XS outputs 52.0 tok/s, and IQ3_XXS outputs 44.8 tok/s.68% 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.68% similarUnverifiedIn the 16-concurrency, 256-token-output test, Gemma 4 E2B achieves an output speed of 1887 tokens/second, and the models faster than it include Qwen2.5-0.5B (4176, 2.21x), Qwen3-0.6B (3210, 1.70x), Qwen2.5-1.5B (2780, 1.47x), and Qwen3-1.7B (2049, 1.09x).66% similarUnverifiedIn the controlled tests with step-by-step improvements, with the configuration of "both run on the GPU alternately", the mean first audio was 2.3 seconds (range 1.9-2.7), and the mean number of stutters per answer was 0.09 (range 0-1), better than "both run on the GPU simultaneously" at 3.9 seconds (3.8-4.1) and 2.84 times (9-17 times).66% similarUnverifiedFor the Qwen2.5-1.5B model, the int4 theoretical roofline ceiling for batch-1 decoding is 411.8 tok/s, while the actual self-developed kernel reaches only 25.5 tok/s, a gap of about 16x.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1092288API
curl https://kongchang.com/api/v1/knowledge/claims/1092288MCP
get_claim(id=1092288)