# LLM Serving Performance on Qwen 7B (Zipf alpha=1.5, 32 variants) Source: [LLM Serving Performance on Qwen 7B (Zipf alpha=1.5, 32 variants)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:06:25.525Z Evaluation of LLM serving performance metrics, including request throughput and end-to-end latency, for various configurations of the Qwen 7B 2.5 model based on Zipf distribution parameters. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | LLM Serving Performance | | Dataset (including any stated split) | Qwen 7B (Zipf distribution, alpha=1.5) 2.5 (32 model variants) | | Metric | Throughput (req/s) | | Unit | SCALAR | | Direction | Higher is better | | Results loaded | 6 of 6 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Throughput (req/s)](): Higher is better; unit: SCALAR. - [E2E Latency (s)](): Lower is better; unit: TIME. ## Recorded results | Method | Throughput (req/s) | Result date | Report paper | | --- | --- | --- | --- | | Delta-CoMe | 0.87 | Not recorded | 68fa7ba31e209131cea8984e | | PRINMIX | 0.87 | Not recorded | 68fa7ba31e209131cea8984e | | Delta-CoMe | 0.42 | Not recorded | 68fa7ba31e209131cea8984e | | PRINMIX | 0.42 | Not recorded | 68fa7ba31e209131cea8984e | | Naive | 0.21 | Not recorded | 68fa7ba31e209131cea8984e | | Naive | 0.18 | Not recorded | 68fa7ba31e209131cea8984e | ### Delta-CoMe Throughput (req/s): 0.87. Result ID: 699a67f45698665ec4f0465d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"arrival rate (lambda)":"1.0"}. This record does not assert independent reproduction. ### PRINMIX Throughput (req/s): 0.87. Result ID: 699a67f45698665ec4f0465f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"arrival rate (lambda)":"1.0"}. This record does not assert independent reproduction. ### Delta-CoMe Throughput (req/s): 0.42. Result ID: 699a67f45698665ec4f0465c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"arrival rate (lambda)":"0.5"}. This record does not assert independent reproduction. ### PRINMIX Throughput (req/s): 0.42. Result ID: 699a67f45698665ec4f0465e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"arrival rate (lambda)":"0.5"}. This record does not assert independent reproduction. ### Naive Throughput (req/s): 0.21. Result ID: 699a67f45698665ec4f0465a. [Reporting paper]() Recorded evaluation configuration: {"arrival rate (lambda)":"0.5"}. This record does not assert independent reproduction. ### Naive Throughput (req/s): 0.18. Result ID: 699a67f45698665ec4f0465b. [Reporting paper]() Recorded evaluation configuration: {"arrival rate (lambda)":"1.0"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.