53.6
% saved ↑
58.8
ttft ms
30.8
Peak memory GB
120
Trials
Validity: valid
The checks LemonMatrix ran while measuring; an invalid run is shown but never ranked.
- ✓ Warm-up discarded
- ✓ Thermal OK
- ✓ Exclusive run
- ✓ No model reload
- TTFT, prefill and decode cover the 75 of 120 requests served locally
- Lemonade does not time cloud requests
- Lemonade doesn't report power draw for this instance; energy metrics omitted
- exclusive_run verified via /api/v1/health polling (no other loaded model was busy)
Trials
Each measured trial after warm-up, from the run's trial sidecar. The summary above is computed from these.
| # | ttft ms | decode tokens per sec | prompt tokens | output tokens | cache tokens | vram gb | host memory gb | route to | matched rule | default used | prompt index | outputs | decode tokens per sec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 49.85 | 49.05 | 15 | 256 | 3 | 30.75 | 47.5 | Qwen3-4B-GGUF | yes | 0 | -- | ||
| 2 | 51.19 | 48.45 | 17 | 256 | 3 | 30.75 | 47.6 | Qwen3-4B-GGUF | yes | 1 | -- | ||
| 3 | 53.24 | 49.2 | 19 | 256 | 3 | 30.75 | 47.6 | Qwen3-4B-GGUF | yes | 2 | -- | ||
| 4 | 48.57 | 49.12 | 18 | 256 | 3 | 30.75 | 47.7 | Qwen3-4B-GGUF | yes | 3 | -- | ||
| 5 | 63.54 | 48.99 | 20 | 256 | 3 | 30.75 | 47.7 | Qwen3-4B-GGUF | yes | 4 | -- | ||
| 6 | 54.79 | 49.01 | 17 | 256 | 3 | 30.75 | 47.7 | Qwen3-4B-GGUF | yes | 5 | -- | ||
| 7 | 50.07 | 49.17 | 14 | 256 | 3 | 30.75 | 47.8 | Qwen3-4B-GGUF | yes | 6 | -- | ||
| 8 | 43.12 | 48.79 | 17 | 256 | 3 | 30.75 | 47.8 | Qwen3-4B-GGUF | yes | 7 | -- | ||
| 9 | 53.69 | 49.29 | 16 | 256 | 3 | 30.75 | 47.8 | Qwen3-4B-GGUF | yes | 8 | -- | ||
| 10 | 45.59 | 49.24 | 16 | 256 | 3 | 30.75 | 47.9 | Qwen3-4B-GGUF | yes | 9 | -- | ||
| 11 | 0 | -- | 29 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 10 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 12 | 0 | -- | 36 | 256 | 0 | 30.75 | 47.8 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 11 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 13 | 0 | -- | 46 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 12 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 14 | 0 | -- | 21 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 13 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 15 | 0 | -- | 30 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 14 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 16 | 0 | -- | 20 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 15 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 17 | 0 | -- | 38 | 256 | 0 | 30.75 | 47.8 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 16 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 18 | 0 | -- | 29 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 17 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 19 | 0 | -- | 19 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 18 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 20 | 0 | -- | 39 | 256 | 0 | 30.75 | 47.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 19 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 21 | 85.4 | 48.95 | 31 | 256 | 3 | 30.75 | 47.9 | Qwen3-4B-GGUF | yes | 20 | -- | ||
| 22 | 73.9 | 48.32 | 27 | 256 | 3 | 30.75 | 47.6 | Qwen3-4B-GGUF | yes | 21 | -- | ||
| 23 | 61.02 | 48.56 | 26 | 256 | 3 | 30.75 | 47.7 | Qwen3-4B-GGUF | yes | 22 | -- | ||
| 24 | 66.22 | 49.14 | 25 | 256 | 3 | 30.75 | 47.8 | Qwen3-4B-GGUF | yes | 23 | -- | ||
| 25 | 62.75 | 48.94 | 25 | 256 | 3 | 30.75 | 47.8 | Qwen3-4B-GGUF | yes | 24 | -- | ||
| 26 | 76.14 | 49.08 | 24 | 256 | 3 | 30.75 | 47.8 | Qwen3-4B-GGUF | yes | 25 | -- | ||
| 27 | 69.73 | 48.28 | 23 | 256 | 3 | 30.75 | 47.9 | Qwen3-4B-GGUF | yes | 26 | -- | ||
| 28 | 50.26 | 48.62 | 22 | 256 | 3 | 30.75 | 48 | Qwen3-4B-GGUF | yes | 27 | -- | ||
| 29 | 63.68 | 48.88 | 23 | 256 | 3 | 30.75 | 48 | Qwen3-4B-GGUF | yes | 28 | -- | ||
| 30 | 52.51 | 48.55 | 22 | 256 | 3 | 30.75 | 48 | Qwen3-4B-GGUF | yes | 29 | -- | ||
| 31 | 96.5 | 47.87 | 185 | 256 | 3 | 30.77 | 48.1 | Qwen3-4B-GGUF | yes | 30 | -- | ||
| 32 | 122.5 | 48.25 | 182 | 256 | 3 | 30.77 | 48.2 | Qwen3-4B-GGUF | yes | 31 | -- | ||
| 33 | 117.8 | 47.78 | 188 | 256 | 3 | 30.77 | 48.2 | Qwen3-4B-GGUF | yes | 32 | -- | ||
| 34 | 101.3 | 47.73 | 185 | 256 | 3 | 30.77 | 48.3 | Qwen3-4B-GGUF | yes | 33 | -- | ||
| 35 | 107.5 | 48.04 | 189 | 256 | 3 | 30.77 | 48.3 | Qwen3-4B-GGUF | yes | 34 | -- | ||
| 36 | 0 | -- | 162 | 256 | 0 | 30.77 | 48.4 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 35 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 37 | 0 | -- | 159 | 256 | 0 | 30.77 | 48.4 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 36 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 38 | 0 | -- | 165 | 256 | 0 | 30.77 | 48.3 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 37 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 39 | 0 | -- | 164 | 256 | 0 | 30.77 | 48.3 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 38 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 40 | 0 | -- | 162 | 182 | 0 | 30.77 | 48.3 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 39 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 41 | 55.12 | 48.99 | 15 | 256 | 3 | 30.77 | 48.4 | Qwen3-4B-GGUF | yes | 0 | -- | ||
| 42 | 55.2 | 48.53 | 17 | 256 | 3 | 30.77 | 48.4 | Qwen3-4B-GGUF | yes | 1 | -- | ||
| 43 | 49.28 | 48.78 | 19 | 256 | 3 | 30.77 | 48.5 | Qwen3-4B-GGUF | yes | 2 | -- | ||
| 44 | 60.23 | 48.58 | 18 | 256 | 3 | 30.77 | 48.5 | Qwen3-4B-GGUF | yes | 3 | -- | ||
| 45 | 52.02 | 48.64 | 20 | 256 | 3 | 30.77 | 48.5 | Qwen3-4B-GGUF | yes | 4 | -- | ||
| 46 | 52.32 | 48.55 | 17 | 256 | 3 | 30.77 | 48.6 | Qwen3-4B-GGUF | yes | 5 | -- | ||
| 47 | 53.77 | 48.44 | 14 | 256 | 3 | 30.77 | 48.6 | Qwen3-4B-GGUF | yes | 6 | -- | ||
| 48 | 58.12 | 48.48 | 17 | 256 | 3 | 30.77 | 48.7 | Qwen3-4B-GGUF | yes | 7 | -- | ||
| 49 | 58.1 | 48.9 | 16 | 256 | 3 | 30.77 | 48.7 | Qwen3-4B-GGUF | yes | 8 | -- | ||
| 50 | 56.5 | 48.43 | 16 | 256 | 3 | 30.77 | 48.7 | Qwen3-4B-GGUF | yes | 9 | -- | ||
| 51 | 0 | -- | 23 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 10 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 52 | 0 | -- | 36 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 11 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 53 | 0 | -- | 52 | 221 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 12 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 54 | 0 | -- | 21 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 13 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 55 | 0 | -- | 24 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 14 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 56 | 0 | -- | 20 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 15 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 57 | 0 | -- | 44 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 16 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 58 | 0 | -- | 29 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 17 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 59 | 0 | -- | 19 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 18 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 60 | 0 | -- | 33 | 256 | 0 | 30.77 | 48.7 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 19 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 61 | 88.91 | 48.63 | 31 | 256 | 3 | 30.77 | 48.7 | Qwen3-4B-GGUF | yes | 20 | -- | ||
| 62 | 59.87 | 48.98 | 27 | 256 | 3 | 30.77 | 48.8 | Qwen3-4B-GGUF | yes | 21 | -- | ||
| 63 | 66.38 | 48.97 | 26 | 256 | 3 | 30.77 | 48.8 | Qwen3-4B-GGUF | yes | 22 | -- | ||
| 64 | 61.16 | 48.87 | 25 | 256 | 3 | 30.77 | 48.9 | Qwen3-4B-GGUF | yes | 23 | -- | ||
| 65 | 75.46 | 49.05 | 25 | 256 | 3 | 30.77 | 49 | Qwen3-4B-GGUF | yes | 24 | -- | ||
| 66 | 67.15 | 48.71 | 24 | 256 | 3 | 30.77 | 49 | Qwen3-4B-GGUF | yes | 25 | -- | ||
| 67 | 70.48 | 48.88 | 23 | 256 | 3 | 30.77 | 49 | Qwen3-4B-GGUF | yes | 26 | -- | ||
| 68 | 64.7 | 48.83 | 22 | 256 | 3 | 30.77 | 49 | Qwen3-4B-GGUF | yes | 27 | -- | ||
| 69 | 50.58 | 49.2 | 23 | 256 | 3 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 28 | -- | ||
| 70 | 68.9 | 48.91 | 22 | 256 | 3 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 29 | -- | ||
| 71 | 25.28 | 47.62 | 185 | 256 | 184 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 30 | -- | ||
| 72 | 24.39 | 48.01 | 182 | 256 | 181 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 31 | -- | ||
| 73 | 24.92 | 47.76 | 188 | 256 | 187 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 32 | -- | ||
| 74 | 24.08 | 47.92 | 185 | 256 | 184 | 30.77 | 49 | Qwen3-4B-GGUF | yes | 33 | -- | ||
| 75 | 24.07 | 48.1 | 189 | 256 | 188 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 34 | -- | ||
| 76 | 0 | -- | 162 | 256 | 0 | 30.77 | 49.1 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 35 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 77 | 0 | -- | 159 | 256 | 64 | 30.77 | 49 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 36 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 78 | 0 | -- | 165 | 256 | 0 | 30.77 | 49.1 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 37 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 79 | 0 | -- | 158 | 256 | 0 | 30.77 | 49.1 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 38 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 80 | 0 | -- | 156 | 256 | 0 | 30.77 | 49.1 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 39 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 81 | 52.92 | 49.02 | 15 | 192 | 3 | 30.77 | 49.1 | Qwen3-4B-GGUF | yes | 0 | -- | ||
| 82 | 57.89 | 48.62 | 17 | 256 | 3 | 30.77 | 49.2 | Qwen3-4B-GGUF | yes | 1 | -- | ||
| 83 | 60.68 | 48.74 | 19 | 256 | 3 | 30.77 | 49.2 | Qwen3-4B-GGUF | yes | 2 | -- | ||
| 84 | 48.37 | 48.6 | 18 | 256 | 3 | 30.77 | 49.3 | Qwen3-4B-GGUF | yes | 3 | -- | ||
| 85 | 65.09 | 49.13 | 20 | 256 | 3 | 30.77 | 49.3 | Qwen3-4B-GGUF | yes | 4 | -- | ||
| 86 | 45.6 | 48.82 | 17 | 256 | 3 | 30.77 | 49.3 | Qwen3-4B-GGUF | yes | 5 | -- | ||
| 87 | 43.29 | 48.8 | 14 | 256 | 3 | 30.77 | 49.4 | Qwen3-4B-GGUF | yes | 6 | -- | ||
| 88 | 59.13 | 48.36 | 17 | 256 | 3 | 30.77 | 49.4 | Qwen3-4B-GGUF | yes | 7 | -- | ||
| 89 | 53.16 | 48.85 | 16 | 256 | 3 | 30.77 | 49.5 | Qwen3-4B-GGUF | yes | 8 | -- | ||
| 90 | 43.9 | 49.1 | 16 | 256 | 3 | 30.77 | 49.5 | Qwen3-4B-GGUF | yes | 9 | -- | ||
| 91 | 0 | -- | 29 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 10 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 92 | 0 | -- | 36 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 11 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 93 | 0 | -- | 46 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 12 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 94 | 0 | -- | 21 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 13 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 95 | 0 | -- | 24 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 14 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 96 | 0 | -- | 20 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 15 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 97 | 0 | -- | 44 | 240 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 16 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 98 | 0 | -- | 23 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 17 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 99 | 0 | -- | 25 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 18 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 100 | 0 | -- | 33 | 256 | 0 | 30.77 | 49.5 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 19 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 101 | 90.43 | 49.53 | 31 | 256 | 3 | 30.77 | 49.6 | Qwen3-4B-GGUF | yes | 20 | -- | ||
| 102 | 78.59 | 48.76 | 27 | 256 | 3 | 30.77 | 49.6 | Qwen3-4B-GGUF | yes | 21 | -- | ||
| 103 | 75.2 | 48.43 | 26 | 256 | 3 | 30.77 | 49.6 | Qwen3-4B-GGUF | yes | 22 | -- | ||
| 104 | 65.12 | 48.46 | 25 | 256 | 3 | 30.77 | 49.7 | Qwen3-4B-GGUF | yes | 23 | -- | ||
| 105 | 64.81 | 48.85 | 25 | 256 | 3 | 30.77 | 49.7 | Qwen3-4B-GGUF | yes | 24 | -- | ||
| 106 | 63.97 | 48.82 | 24 | 256 | 3 | 30.77 | 49.8 | Qwen3-4B-GGUF | yes | 25 | -- | ||
| 107 | 69.17 | 48.83 | 23 | 256 | 3 | 30.77 | 49.8 | Qwen3-4B-GGUF | yes | 26 | -- | ||
| 108 | 54.31 | 48.67 | 22 | 256 | 3 | 30.77 | 49.9 | Qwen3-4B-GGUF | yes | 27 | -- | ||
| 109 | 72.67 | 48.64 | 23 | 256 | 3 | 30.77 | 49.9 | Qwen3-4B-GGUF | yes | 28 | -- | ||
| 110 | 57.49 | 48.59 | 22 | 256 | 3 | 30.77 | 50 | Qwen3-4B-GGUF | yes | 29 | -- | ||
| 111 | 24.16 | 47.95 | 185 | 256 | 184 | 30.77 | 49.9 | Qwen3-4B-GGUF | yes | 30 | -- | ||
| 112 | 25.26 | 47.62 | 182 | 256 | 181 | 30.77 | 49.9 | Qwen3-4B-GGUF | yes | 31 | -- | ||
| 113 | 23.91 | 47.5 | 188 | 256 | 187 | 30.77 | 50 | Qwen3-4B-GGUF | yes | 32 | -- | ||
| 114 | 23.3 | 47.87 | 185 | 256 | 184 | 30.77 | 50 | Qwen3-4B-GGUF | yes | 33 | -- | ||
| 115 | 23.66 | 47.6 | 189 | 256 | 188 | 30.77 | 50 | Qwen3-4B-GGUF | yes | 34 | -- | ||
| 116 | 0 | -- | 162 | 256 | 0 | 30.77 | 49.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 35 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 117 | 0 | -- | 165 | 256 | 0 | 30.77 | 49.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 36 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 118 | 0 | -- | 165 | 256 | 0 | 30.77 | 49.8 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 37 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 119 | 0 | -- | 164 | 256 | 0 | 30.77 | 49.8 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 38 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} | |
| 120 | 0 | -- | 162 | 174 | 0 | 30.77 | 49.9 | openrouter.deepseek/deepseek-v3.2 | code-to-cloud | no | 39 | {"estimated_cost": {"cost_input_per_million": 0.26, "cost_output_per_million": 0.38}} |
Model and configuration
| model.name | user.LM-A-keyword-code-to-cloud |
| model.class | router |
| model.quantization | none |
| model.context_length | 4096 |
| config.compute_engine | router |
| config.backend | collection.router |
| config.os | linux |
| config.power_state | plugged |
| config.router_default_model | Qwen3-4B-GGUF |
Environment
| cpu | AMD Ryzen 9 9950X3D 16-Core Processor |
| device_model | AMD Ryzen 9 9950X3D 16-Core Processor |
| dgpu | AMD Radeon AI PRO R9700 (gfx1201) |
| dgpu_family | gfx1201 |
| driver_version | unknown |
| igpu | AMD Ryzen 9 9950X3D 16-Core Processor (gfx1036) |
| igpu_family | gfx1036 |
| lemonade_version | 2026.46.2 |
| memory_gb | 59.46 |
| os_version | Linux-7.0.0-42-generic (Ubuntu 26.04) |
All metrics
| prefill.tokens_per_sec | 915.7 |
| prefill.stddev | 2,490.2 |
| prefill.p95 | 7,733.7 |
| prefill.ci95_half_width | 563.6 |
| decode.tokens_per_sec | 48.62 |
| decode.inter_token_latency_ms | 20.57 |
| decode.stddev | 0.468 |
| decode.p95 | 49.2 |
| decode.ci95_half_width | 0.1059 |
| ttft_ms | 58.8 |
| trial_count | 120 |
| peak_memory_gb | 30.77 |
| prompt_sha256 | 5061664e3d07a0e9e1e54868eddcbf4da9b5ccbf0651c94994b016bf595d30a8 |
| routing.requests | 120 |
| routing.candidates.Qwen3-4B-GGUF.requests | 75 |
| routing.candidates.Qwen3-4B-GGUF.prompt_tokens | 4038 |
| routing.candidates.Qwen3-4B-GGUF.output_tokens | 19136 |
| routing.candidates.Qwen3-4B-GGUF.cost_usd | 0 |
| routing.candidates.Qwen3-4B-GGUF.price_source | local_default |
| routing.candidates.Qwen3-4B-GGUF.share | 0.625 |
| routing.candidates.openrouter.deepseek/deepseek-v3.2.requests | 45 |
| routing.candidates.openrouter.deepseek/deepseek-v3.2.prompt_tokens | 3339 |
| routing.candidates.openrouter.deepseek/deepseek-v3.2.output_tokens | 11313 |
| routing.candidates.openrouter.deepseek/deepseek-v3.2.cost_usd | 0.005167 |
| routing.candidates.openrouter.deepseek/deepseek-v3.2.price_source | override |
| routing.candidates.openrouter.deepseek/deepseek-v3.2.share | 0.375 |
| routing.local_share | 0.625 |
| routing.free_share | 0.625 |
| routing.routed_cost_usd | 0.005167 |
| routing.unpriced_requests | 0 |
| routing.prompt_set_size | 40 |
| routing.baseline_model | openrouter.deepseek/deepseek-v3.2 |
| routing.baseline_method | measured |
| routing.baseline_estimated_requests | 0 |
| routing.baseline_cost_usd | 0.01114 |
| routing.savings_usd | 0.005972 |
| routing.savings_pct | 53.61 |
| ttft_ms_stddev | 20.88 |
| ttft_ms_ci95_half_width | 4.727 |