1{2 "timestamp_utc": "2026-10-30T03:22:24.245610Z",3 "models": {4 "qwen3b": {5 "repo_id": "mlx-community/Qwen2.5-3B-Instruct-bf16",6 "snapshot_path": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--Qwen2.5-3B-Instruct-bf16/snapshots/bec1ad658e380044bbfe06d4a2268801d3379a11",7 "revision": "bec1ad658e380044bbfe06d4a2268801d3379a11",8 "model_type": "qwen2",9 "num_layers": 36,10 "num_kv_heads": 2,11 "num_attention_heads": 16,12 "head_dim": 128,13 "hidden_size": 2048,14 "vocab_size": 151936,15 "max_position_embeddings": 32768,16 "rope_theta": 1000000.0,17 "rope_scaling": null,18 "sliding_window": 32768,19 "weight_dtype": "bfloat16",20 "weight_quantization": null,21 "tokenizer_hash": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b",22 "chat_template_hash": "d5495a1e5db0611132a97e46a65dbb64a642a499421228b9c8b93229097fa9a4",23 "config_hash": "9f4d0bfcc2df00e16cd130a48c908ab5c5a982d9c656c670f072ecd0d034010f",24 "weight_files_sha256": {25 "model-00001-of-00002.safetensors": "feb3c8c4721f4503d07509e59c1768fe90804340f1d28d9c2aa406215eb6490b",26 "model-00002-of-00002.safetensors": "77b5846eb11fecb6eb27fb70bcc0d0c73a57b8443d01ebcaa558f4066eb34b70"27 },28 "mlx_version": "0.32.4",29 "mlx_lm_version": "0.32.2",30 "bos_token_id": null,31 "eos_token_ids": [32 15164533 ],34 "probe_answer": "ZR-77Q-0931",35 "probe_ok": true,36 "load_ms": 852.6365419384092,37 "context_lengths": {38 "2048": {39 "supported": true,40 "prefill_ms": 528.9326659403741,41 "cache_bytes": 75497472,42 "peak_bytes": 705478660443 },44 "8192": {45 "supported": true,46 "prefill_ms": 967.8050000220537,47 "cache_bytes": 301989888,48 "peak_bytes": 732324665649 },50 "16384": {51 "supported": true,52 "prefill_ms": 2285.5356669751927,53 "cache_bytes": 603979776,54 "peak_bytes": 749105158455 }56 },57 "memory_after_probe": {58 "mlx_active_bytes": 6171877384,59 "mlx_peak_bytes": 7491051584,60 "mlx_cache_bytes": 3520642368,61 "process_rss_bytes": 660340736062 }63 },64 "llama3b": {65 "repo_id": "mlx-community/Llama-3.2-3B-Instruct-bf16",66 "snapshot_path": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--Llama-3.2-3B-Instruct-bf16/snapshots/6d88ba43024fef71b10e52e101c7cd4598322601",67 "revision": "6d88ba43024fef71b10e52e101c7cd4598322601",68 "model_type": "llama",69 "num_layers": 28,70 "num_kv_heads": 8,71 "num_attention_heads": 24,72 "head_dim": 128,73 "hidden_size": 3072,74 "vocab_size": 128256,75 "max_position_embeddings": 131072,76 "rope_theta": 500000.0,77 "rope_scaling": {78 "factor": 32.0,79 "high_freq_factor": 4.0,80 "low_freq_factor": 1.0,81 "original_max_position_embeddings": 8192,82 "rope_type": "llama3"83 },84 "sliding_window": null,85 "weight_dtype": "bfloat16",86 "weight_quantization": null,87 "tokenizer_hash": "b9d8872cf3a99da99e1fc4b0f71232a451632dc49ca4c6771549b4efa5c1c763",88 "chat_template_hash": "5816fce10444e03c2e9ee1ef8a4a1ea61ae7e69e438613f3b17b69d0426223a4",89 "config_hash": "2b13c35350c57b0c3130063f944cd3a4353c0463fb664e80172d479fb451f923",90 "weight_files_sha256": {91 "model-00001-of-00002.safetensors": "b475af8535933afaf9393b30c6bcb4bfede59586b16ecdd010e7da8a83dfc95b",92 "model-00002-of-00002.safetensors": "9b2903fa2b3728423f6442c0dc667db645906fb113c9d0cd794faeb820cc82d3"93 },94 "mlx_version": "0.32.4",95 "mlx_lm_version": "0.32.2",96 "bos_token_id": 128000,97 "eos_token_ids": [98 128008,99 128001,100 128009101 ],102 "probe_answer": "ZR-77Q-0931",103 "probe_ok": true,104 "load_ms": 2813.813167042099,105 "context_lengths": {106 "2048": {107 "supported": true,108 "prefill_ms": 251.04662496596575,109 "cache_bytes": 234881024,110 "peak_bytes": 7142741028111 },112 "8192": {113 "supported": true,114 "prefill_ms": 1006.7105001071468,115 "cache_bytes": 939524096,116 "peak_bytes": 7851602992117 },118 "16384": {119 "supported": true,120 "prefill_ms": 2449.0344589576125,121 "cache_bytes": 1879048192,122 "peak_bytes": 8778576936123 }124 },125 "memory_after_probe": {126 "mlx_active_bytes": 6425506824,127 "mlx_peak_bytes": 8778576936,128 "mlx_cache_bytes": 8951880172,129 "process_rss_bytes": 6437617664130 }131 },132 "qwen1b5": {133 "repo_id": "mlx-community/Qwen2.5-1.5B-Instruct-bf16",134 "snapshot_path": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--Qwen2.5-1.5B-Instruct-bf16/snapshots/4ae77cb209f06199b8df1c94e21ff341332a3a89",135 "revision": "4ae77cb209f06199b8df1c94e21ff341332a3a89",136 "model_type": "qwen2",137 "num_layers": 28,138 "num_kv_heads": 2,139 "num_attention_heads": 12,140 "head_dim": 128,141 "hidden_size": 1536,142 "vocab_size": 151936,143 "max_position_embeddings": 32768,144 "rope_theta": 1000000.0,145 "rope_scaling": null,146 "sliding_window": 32768,147 "weight_dtype": "bfloat16",148 "weight_quantization": null,149 "tokenizer_hash": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b",150 "chat_template_hash": "d5495a1e5db0611132a97e46a65dbb64a642a499421228b9c8b93229097fa9a4",151 "config_hash": "a1cd2dc97c13a2dce95741f912fa212cfa321260c42627202a122b80c68a52c9",152 "weight_files_sha256": {153 "model.safetensors": "073400b311d049b6379b8f3f7ad5a0258268031dd009b99389cfd4030144850a"154 },155 "mlx_version": "0.32.4",156 "mlx_lm_version": "0.32.2",157 "bos_token_id": null,158 "eos_token_ids": [159 151645160 ],161 "probe_answer": "ZR-77Q-0931",162 "probe_ok": true,163 "load_ms": 500.1337080029771,164 "context_lengths": {165 "2048": {166 "supported": true,167 "prefill_ms": 103.63012494053692,168 "cache_bytes": 58720256,169 "peak_bytes": 3775302700170 },171 "8192": {172 "supported": true,173 "prefill_ms": 493.9766660099849,174 "cache_bytes": 234881024,175 "peak_bytes": 3993431104176 },177 "16384": {178 "supported": true,179 "prefill_ms": 1201.0092079872265,180 "cache_bytes": 469762048,181 "peak_bytes": 4261899328182 }183 },184 "memory_after_probe": {185 "mlx_active_bytes": 3087428616,186 "mlx_peak_bytes": 4261899328,187 "mlx_cache_bytes": 2816668640,188 "process_rss_bytes": 3474096128189 }190 }191 }192}