{"verdict":"yes","receipt":"https://fitllm.run/r/glm-4-7-flash-q4_k_m-on-rtx-3090","model":"GLM-4.7-Flash","hardware":"RTX 3090","env":"windows-display","quant":"Q4_K_M","kvBits":16,"ctx":8192,"structuralAssumptions":[{"id":"mla-compressed-latent-cache","statement":"KV memory assumes a compressed-latent MLA artifact or mode; legacy non-MLA GGUF or an explicitly uncompressed mode invalidates this estimate."}],"memory":{"usedGB":22.63,"totalGB":24,"freeGB":1.37,"breakdown":{"weightsGB":17.78,"kvCacheGB":0.41,"overheadGB":2.44,"reserveGB":2}},"maxContext":10109,"contextLimits":{"recommendedMax":10109,"hardLimit":23747,"headroomPct":5},"fix":null,"note":"usedGB = total required to run: weights + KV cache + runtime overhead + OS/GPU reserve. Runtime resident-weights measurements (e.g. oMLX actual_size) read lower — a different metric, not a prediction error.","source":"https://fitllm.run","engine":"fitllm-engine (MIT)","data":"https://fitllm.run/data (CC0)"}