| { | |
| "model": ".cache/Qwen3.5-0.8B", | |
| "prompt": "Describe a queue in one short sentence.", | |
| "nonstream": "A queue is a line waiting for service.\n", | |
| "streamed": "A queue is a line waiting for service.\n", | |
| "stream_matches_nonstream": true, | |
| "chunks": 8, | |
| "input_tokens": 20, | |
| "output_tokens_retokenized": 10, | |
| "seconds": 3.7454539999598637, | |
| "time_to_first_text_seconds": 1.1324454999994487, | |
| "note": "Text chunks can buffer multiple tokens; first text is not exact first-token latency", | |
| "limitations": "One greedy request, warm model, no concurrent load or audio" | |
| } |