{"model":{"id":"qwen3-32b","name":"Qwen3-32B"},"hardware":{"id":"rtx-4090","name":"GeForce RTX 4090","memoryGB":24},"contextTokens":8192,"bestFit":{"quant":"Q4_K_M","fits":true,"weightsGiB":18.5,"kvCacheGiB":2,"overheadGiB":1.4,"totalGiB":21.9,"usableGiB":24,"estimatedTokensPerSecond":35.5},"byQuant":[{"quant":"Q4_K_M","fits":true,"weightsGiB":18.5,"kvCacheGiB":2,"overheadGiB":1.4,"totalGiB":21.9,"usableGiB":24,"estimatedTokensPerSecond":35.5},{"quant":"Q5_K_M","fits":false,"weightsGiB":21.7,"kvCacheGiB":2,"overheadGiB":1.6,"totalGiB":25.3,"usableGiB":24,"estimatedTokensPerSecond":null},{"quant":"Q6_K","fits":false,"weightsGiB":25.2,"kvCacheGiB":2,"overheadGiB":1.8,"totalGiB":28.9,"usableGiB":24,"estimatedTokensPerSecond":null},{"quant":"Q8_0","fits":false,"weightsGiB":32.5,"kvCacheGiB":2,"overheadGiB":2.1,"totalGiB":36.6,"usableGiB":24,"estimatedTokensPerSecond":null},{"quant":"FP16","fits":false,"weightsGiB":61.1,"kvCacheGiB":2,"overheadGiB":3.6,"totalGiB":66.6,"usableGiB":24,"estimatedTokensPerSecond":null}],"note":"Estimates, not benchmarks. Method: https://onpremworld.com/en/methodology","updated":"2026-10-14","license":"CC BY 4.0","source":"https://onpremworld.com","docs":"https://onpremworld.com/openapi.json"}