{ "generated": "2026-07-18T12:30:55", "modes": [ "cpu", "gpu0:NVIDIA_GeForce_RTX_3090", "gpu1:NVIDIA_RTX_A4000", "gpu2:NVIDIA_RTX_A4000" ], "results": [ { "file": "Yi-Coder-9B-Chat-Q4_K_M.gguf", "size_mb": 5082.1, "cpu": { "load_seconds": 2.4, "gen_tps_avg": 7.8, "prefill_tps_avg": 32.1 }, "gpu0:NVIDIA_GeForce_RTX_3090": { "load_seconds": 2.0, "gen_tps_avg": 120.1, "prefill_tps_avg": 681.2 }, "gpu1:NVIDIA_RTX_A4000": { "load_seconds": 2.0, "gen_tps_avg": 64.2, "prefill_tps_avg": 490.9 }, "gpu2:NVIDIA_RTX_A4000": { "load_seconds": 2.0, "gen_tps_avg": 64.8, "prefill_tps_avg": 493.6 }, "gen_tps_avg": 120.1, "load_seconds": 2.4, "prefill_tps_avg": 32.1, "gpu_speedup": 15.4 }, { "file": "Yi-Coder-9B-Chat-Q5_K_M.gguf", "size_mb": 5968.3, "cpu": { "load_seconds": 2.4, "gen_tps_avg": 6.8, "prefill_tps_avg": 27.8 }, "gpu0:NVIDIA_GeForce_RTX_3090": { "load_seconds": 2.0, "gen_tps_avg": 107.9, "prefill_tps_avg": 646.9 }, "gpu1:NVIDIA_RTX_A4000": { "load_seconds": 3.2, "gen_tps_avg": 56.4, "prefill_tps_avg": 449.2 }, "gpu2:NVIDIA_RTX_A4000": { "load_seconds": 2.0, "gen_tps_avg": 56.9, "prefill_tps_avg": 470.7 }, "gen_tps_avg": 107.9, "load_seconds": 2.4, "prefill_tps_avg": 27.8, "gpu_speedup": 15.9 }, { "file": "Yi-Coder-9B-Chat-Q6_K.gguf", "size_mb": 6910.0, "cpu": { "load_seconds": 2.5, "gen_tps_avg": 5.9, "prefill_tps_avg": 24.1 }, "gpu0:NVIDIA_GeForce_RTX_3090": { "load_seconds": 3.1, "gen_tps_avg": 93.2, "prefill_tps_avg": 634.9 }, "gpu1:NVIDIA_RTX_A4000": { "load_seconds": 3.1, "gen_tps_avg": 47.1, "prefill_tps_avg": 451.8 }, "gpu2:NVIDIA_RTX_A4000": { "load_seconds": 3.1, "gen_tps_avg": 48.6, "prefill_tps_avg": 452.2 }, "gen_tps_avg": 93.2, "load_seconds": 2.5, "prefill_tps_avg": 24.1, "gpu_speedup": 15.8 }, { "file": "Yi-Coder-9B-Chat-Q8_0.gguf", "size_mb": 8949.2, "cpu": { "load_seconds": 2.7, "gen_tps_avg": 4.8, "prefill_tps_avg": 18.7 }, "gpu0:NVIDIA_GeForce_RTX_3090": { "load_seconds": 3.2, "gen_tps_avg": 80.5, "prefill_tps_avg": 662.1 }, "gpu1:NVIDIA_RTX_A4000": { "load_seconds": 3.2, "gen_tps_avg": 40.2, "prefill_tps_avg": 447.2 }, "gpu2:NVIDIA_RTX_A4000": { "load_seconds": 3.6, "gen_tps_avg": 40.3, "prefill_tps_avg": 437.5 }, "gen_tps_avg": 80.5, "load_seconds": 2.7, "prefill_tps_avg": 18.7, "gpu_speedup": 16.8 } ] }