community pool / model answer

Qwen2.5-Coder-14B-Instruct

code · Qwen/Qwen2.5-Coder-14B-Instruct

35 tok/s · model median · 13 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
RTX 2080 8GB4-bit11 tok/sreported-9 GB2,048ollama
RTX 3060 12GB4-bit35 tok/sreported-9 GB2,048ollama
RTX 3080 12GB4-bit72 tok/smeasured-9 GB2,048llama.cpp, ollama
RTX 5000 16GB5-bit24.3 tok/sreported-14.6 GB32,768llama.cpp
RTX PRO 6000 Blackwell 96GB16-bit50.3 tok/sreported-- GB32,768vllm
RTX3060+RTX2080-pooled-20GB 20GB4-bit35 tok/smeasured-9 GB2,048ollama

Context tested

The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.