RTX 4090
110 records across 17 models (quants: IQ3_XS, MXFP4, Q4_0, Q4_K, Q4_K_M, Q4_K_S, Q4_K_XL, Ternary PQ2_0, Ternary PTQ1_0, UD-Q4_K_XL).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RTX 4090 | GLM-4.7-Flash | Q4_K | llama.cpp (CUDA) | – | 6750.5 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | GLM-4.7-Flash | Q4_K | llama.cpp (CUDA) | – | 6843.1 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | GLM-4.7-Flash | Q4_K | llama.cpp (CUDA) | – | 5732.7 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | GLM-4.7-Flash | Q4_K | llama.cpp (CUDA) | – | 5794.6 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | GLM-4.7-Flash | Q4_K | llama.cpp (CUDA) | – | 8437.7 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | GLM-4.7-Flash | Q4_K | llama.cpp (CUDA) | – | 8587.5 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | gpt-oss 20B 20B | MXFP4 | llama.cpp | 190.6 | 8369.3 | – | – | 4096 | – | sourced | – | source |
| RTX 4090 | gpt-oss 20B 20B | MXFP4 | llama.cpp | 163.3 | 5856.2 | – | – | 16384 | – | sourced | – | source |
| RTX 4090 | gpt-oss 20B 20B | MXFP4 | llama.cpp | 140.7 | 3705.6 | – | – | 32768 | – | sourced | – | source |
| RTX 4090 | gpt-oss 20B 20B | MXFP4 | llama.cpp | 104.0 | 2092.2 | – | – | 65536 | – | sourced | – | source |
| RTX 4090 | gpt-oss 20B 20B | MXFP4 | llama.cpp | 70.3 | 1114.7 | – | – | 131072 | – | sourced | – | source |
| RTX 4090 | Kimi-Linear-48B-A3B | IQ3_XS | llama.cpp (CUDA) | – | 5861.0 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Kimi-Linear-48B-A3B | IQ3_XS | llama.cpp (CUDA) | – | 6041.0 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Kimi-Linear-48B-A3B | IQ3_XS | llama.cpp (CUDA) | – | 5854.8 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Kimi-Linear-48B-A3B | IQ3_XS | llama.cpp (CUDA) | – | 6022.7 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 150.59 | 10112.60 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 150.4 | 9365 | – | – | tg128 | 2026-04-03 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 150.17 | 10076.76 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 149.3 | 10060 | – | – | tg128 | 2026-04-03 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 140.0 | 10903 | – | – | tg128 | 2026-04-03 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 139.83 | 10957.29 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 138.4 | 10925 | – | – | tg128 | 2026-04-03 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 137.81 | 10936.64 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 136.37 | 10808.36 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 134.73 | 10794.80 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_XL | llama.cpp llama-bench | 131 | – | – | – | 4096 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 129.26 | 10571.92 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 128.3 | 10476 | – | – | tg128 | 2026-04-03 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 127.79 | 10456.24 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 124.3 | 10405 | – | – | tg128 | 2026-04-03 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 123.95 | 10384.88 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_M | llama.cpp (CUDA) | 122.68 | 10154.79 | – | – | tg128 | 2026-04-04 | community | – | source |
| RTX 4090 | Llama 3.1 8B | Q4_K_XL | llama.cpp llama-bench | 53.07 | – | – | – | 65536 | 2026-08-17 | sourced | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 190.1 | 10830.41 | – | – | tg128 | 2025-09-24 | community | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (CUDA) | 189.07 | 14753.2 | – | – | tg128 | 2025-08-03 | community | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 187.97 | 9452.03 | – | – | tg128 | 2025-09-24 | community | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (CUDA) | 185.04 | 14472.84 | – | – | tg128 | 2025-07-08 | community | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 179.62 | 9481.99 | – | – | tg128 | 2025-08-03 | community | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 170.95 | 8534.56 | – | – | tg128 | 2025-07-08 | community | – | source |
| RTX 4090 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 166.45 | 9032.44 | – | – | tg128 | 2025-07-08 | community | – | source |
| RTX 4090 | Ornith-1.5-35B | Q4_K_M | llama.cpp (CUDA) | 113.8 | 2210 | – | – | 160000 | 2026-10-07 | community | – | source |
| RTX 4090 | qwen-2.5-7b | Q4_K_M | llama.cpp llama-bench | 135 | – | – | – | 4096 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen2.5-7B | Q4_0 | llama.cpp (CUDA) | 163.51 | – | – | – | 4096 | 2026-10-05 | community | – | source |
| RTX 4090 | Qwen2.5-7B | Q4_0 | llama.cpp (CUDA) | 149.39 | – | – | – | 4096 | 2026-10-05 | community | – | source |
| RTX 4090 | Qwen2.5-7B | Q4_0 | llama.cpp (CUDA) | 136.26 | – | – | – | 16384 | 2026-10-05 | community | – | source |
| RTX 4090 | Qwen2.5-7B | Q4_0 | llama.cpp (CUDA) | 110.98 | – | – | – | 32768 | 2026-10-05 | community | – | source |
| RTX 4090 | Qwen2.5-7B | Q4_0 | llama.cpp (CUDA) | 110.87 | – | – | – | 16384 | 2026-10-05 | community | – | source |
| RTX 4090 | Qwen2.5-7B | Q4_0 | llama.cpp (CUDA) | 88.07 | – | – | – | 32768 | 2026-10-05 | community | – | source |
| RTX 4090 | Qwen3 14B 14B | Q4_K | llama.cpp | 84.4 | 5265.4 | – | – | 4096 | – | sourced | – | source |
| RTX 4090 | Qwen3 14B | Q4_K_XL | llama.cpp llama-bench | 82.82 | – | – | – | 4096 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 14B 14B | Q4_K | llama.cpp | 69.8 | 3452.3 | – | – | 16384 | – | sourced | – | source |
| RTX 4090 | Qwen3 14B 14B | Q4_K | llama.cpp | 55.4 | 2324.2 | – | – | 32768 | – | sourced | – | source |
| RTX 4090 | Qwen3 14B 14B | Q4_K | llama.cpp | 38.8 | 1398.3 | – | – | 65536 | – | sourced | – | source |
| RTX 4090 | Qwen3 14B | Q4_K_XL | llama.cpp llama-bench | 38.73 | – | – | – | 65536 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 30B A3B | Q4_K_M | llama.cpp (CUDA) | 225.9 | 1130.8 | – | – | tg128 | 2026-04-01 | community | – | source |
| RTX 4090 | Qwen3 30B A3B | Q4_K_M | llama.cpp (CUDA) | 223.2 | 1132.8 | – | – | tg128 | 2026-04-01 | community | – | source |
| RTX 4090 | Qwen3 30B A3B | Q4_K_M | llama.cpp (CUDA) | 211.0 | 1087.8 | – | – | tg128 | 2026-04-01 | community | – | source |
| RTX 4090 | Qwen3 30B A3B 30B | Q4_K | llama.cpp | 207.3 | 6159.4 | – | – | 4096 | – | sourced | – | source |
| RTX 4090 | Qwen3 30B A3B | Q4_K_XL | llama.cpp llama-bench | 195.78 | – | – | – | 4096 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 30B A3B 30B | Q4_K | llama.cpp | 139.2 | 4027.3 | – | – | 16384 | – | sourced | – | source |
| RTX 4090 | Qwen3 30B A3B 30B | Q4_K | llama.cpp | 105.1 | 2627.7 | – | – | 32768 | – | sourced | – | source |
| RTX 4090 | Qwen3 30B A3B | Q4_K_XL | llama.cpp llama-bench | 74.64 | – | – | – | 57344 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 30B A3B 30B | Q4_K | llama.cpp | 68.2 | 1502.5 | – | – | 65536 | – | sourced | – | source |
| RTX 4090 | Qwen3 32B 32B | Q4_K | llama.cpp | 39.6 | 2448.0 | – | – | 4096 | – | sourced | – | source |
| RTX 4090 | Qwen3 32B 32B | Q4_K | llama.cpp | 34.4 | 1707.4 | – | – | 16384 | – | sourced | – | source |
| RTX 4090 | Qwen3 8B 8B | Q4_K | llama.cpp | 141.3 | 9250.5 | – | – | 4096 | – | sourced | – | source |
| RTX 4090 | Qwen3 8B | Q4_K_M | llama.cpp llama-bench | 141.3 | – | – | – | 4096 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 8B 8B | Q4_K | llama.cpp | 108.0 | 5530.5 | – | – | 16384 | – | sourced | – | source |
| RTX 4090 | Qwen3 8B | Q4_K_M | llama.cpp llama-bench | 108 | – | – | – | 16384 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 8B 8B | Q4_K | llama.cpp | 82.3 | 3560.1 | – | – | 32768 | – | sourced | – | source |
| RTX 4090 | Qwen3 8B | Q4_K_M | llama.cpp llama-bench | 82.3 | – | – | – | 32768 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 8B 8B | Q4_K | llama.cpp | 56.1 | 2028.6 | – | – | 65536 | – | sourced | – | source |
| RTX 4090 | Qwen3 8B | Q4_K_M | llama.cpp llama-bench | 56.1 | – | – | – | 65536 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3 8B 8B | Q4_K | llama.cpp | 33.8 | 1059.9 | – | – | 131072 | – | sourced | – | source |
| RTX 4090 | Qwen3 8B | Q4_K_M | llama.cpp llama-bench | 33.8 | – | – | – | 131072 | 2026-08-17 | sourced | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp (CUDA) | – | 7473.5 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp (CUDA) | – | 7570.8 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp (CUDA) | – | 7283.5 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp (CUDA) | – | 7370.4 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp (CUDA) | – | 11554.2 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp (CUDA) | – | 11665.8 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp | – | 10872.34 | – | – | 2048 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp | – | 12044.25 | – | – | 2048 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp | – | 11187.34 | – | – | 4096 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.5-35B-A3B | Q4_K_S | llama.cpp | – | 12428.05 | – | – | 4096 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.5-9B | Q4_K_M | llama.cpp (CUDA) | – | 9343.1 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-9B | Q4_K_M | llama.cpp (CUDA) | – | 9515.0 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-9B | Q4_K_M | llama.cpp (CUDA) | – | 9472.7 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.5-9B | Q4_K_M | llama.cpp (CUDA) | – | 9619.7 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | 47.67 | – | – | – | 128 | 2026-10-07 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | 47.62 | – | – | – | 128 | 2026-10-07 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3075.24 | – | – | 512 | 2026-10-07 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3243.6 | – | – | 512 | 2026-10-07 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3059.87 | – | – | 4096 | 2026-10-07 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3221.08 | – | – | 4096 | 2026-10-07 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 2984.8 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3057.3 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 2954.5 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3030.0 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 2950.6 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | UD-Q4_K_XL | llama.cpp (CUDA) | – | 3000.7 | – | – | – | 2026-10 | community | – | source |
| RTX 4090 | Qwen3.8-27B | Q4_K_M | llama.cpp | – | 3336.69 | – | – | 2048 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.8-27B | Q4_K_M | llama.cpp | – | 3438.65 | – | – | 2048 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.8-27B | Q4_K_M | llama.cpp | – | 3340.9 | – | – | 4096 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.8-27B | Q4_K_M | llama.cpp | – | 3446.88 | – | – | 4096 | 2026-10-13 | sourced | – | source |
| RTX 4090 | Qwen3.8-Flash-Next | Q4_0 | llama.cpp (CUDA) | 40.7 | – | – | – | tg512 | 2026-10-03 | community | – | source |
| RTX 4090 | Qwen3.8-Flash-Next | Q4_0 | llama.cpp (CUDA) | 39.4 | – | – | – | tg512 | 2026-10-03 | community | – | source |
| RTX 4090 | Qwen3.8-Flash-Next | Q4_0 | llama.cpp (CUDA) | 25.0 | – | – | – | tg512 | 2026-10-03 | community | – | source |
| RTX 4090 | Ternary-Bonsai-2-27B | Ternary PTQ1_0 | llama.cpp (CUDA) | 91.1 | 1645 | – | – | tg128 | 2026-09-16 | community | – | source |
| RTX 4090 | Ternary-Bonsai-2-27B | Ternary PQ2_0 | llama.cpp (CUDA) | 81.2 | 3124 | – | – | tg128 | 2026-09-16 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085