{ "schemaVersion": 1, "snapshotId": "sha256-5858f01d391677077ef493c23f47bc702284124050a3bf0f76b052acc36a5a57", "dataAsOf": "2026-10-15", "defaults": { "read": 0.9, "write": 0.1, "hours": 24, "electricity": 0.15, "life": 5, "apiInput": 0.14, "apiOutput": 0.28 }, "records": [ { "recordKey": "dual5070:qwen38_27b", "rigId": "dual5070", "hardware": "2× RTX 5070 Ti", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 32, "readingTps": 2238.15, "writingTps": 97.04, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 13, "serviceObservationCount": 0, "asOf": "2026-08-31", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/unsloth/Qwen3.8-27B-NVFP4/discussions/16", "https://huggingface.co/eemin/Qwen3.8-27B-NVFP4-MTP-2x16GB" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macmini_m5pro_64:qwen38_27b", "rigId": "macmini_m5pro_64", "hardware": "Mac mini M5 Pro 64 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 64, "readingTps": 1482.75, "writingTps": 36.775, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 5, "serviceObservationCount": 0, "asOf": "2026-08-19", "seoEligible": true, "sourceUrls": [ "https://omlx.ai/benchmarks/performance/5jfk11m0", "https://omlx.ai/benchmarks/performance/e5l8yvgy", "https://omlx.ai/benchmarks/performance/126q6slv", "https://omlx.ai/benchmarks/performance/d5k5l4tv" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtxpro5000:qwen38_27b", "rigId": "rtxpro5000", "hardware": "RTX PRO 5000 Blackwell", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 48, "readingTps": 6288, "writingTps": 112.5, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-08-23", "seoEligible": true, "sourceUrls": [ "https://github.com/vllm-project/vllm/issues/53428" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx5070:qwen38_27b", "rigId": "rtx5070", "hardware": "RTX 5070 Ti", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 16, "readingTps": 1555.5555555555557, "writingTps": 50.8, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-08-18", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5max_64:qwen38_27b", "rigId": "macstudio_m5max_64", "hardware": "Mac Studio M5 Max 64 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 64, "readingTps": 745, "writingTps": 59.2, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": "2026-08-15", "seoEligible": true, "sourceUrls": [ "https://www.apple.com/shop/buy-mac/mac-studio/m5-max-chip-18-core-cpu-40-core-gpu-64gb-memory-1tb-storage" ], "estimateSource": "Projection anchored to the reported M5 Max Qwen3.8 source median; no Mac Studio measurement yet", "estimateAnchors": [ { "factor": 1, "rig": "m5max", "sourceBandwidth": 614, "targetBandwidth": 614 } ], "projectionMethod": null }, { "recordKey": "rtx3090ti:qwen38_27b", "rigId": "rtx3090ti", "hardware": "RTX 3090 Ti", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 24, "readingTps": 1458.3333333333335, "writingTps": 69.75, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-08-18", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx4090:qwen38_27b", "rigId": "rtx4090", "hardware": "RTX 4090", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 24, "readingTps": 2692.3076923076924, "writingTps": 65.275, "readingStatus": "measured", "writingStatus": "measured", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 6, "serviceObservationCount": 0, "asOf": "2026-08-24", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24", "https://github.com/sudoingX/qwen38-mtp" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx5090:qwen38_27b", "rigId": "rtx5090", "hardware": "RTX 5090", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 32, "readingTps": 3571.4285714285716, "writingTps": 111.8, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 13, "serviceObservationCount": 0, "asOf": "2026-08-24", "seoEligible": true, "sourceUrls": [ "https://github.com/sudoingX/qwen38-mtp", "https://www.localmaxxing.com/api/speed-tests?username=CsGoat&date=2026-08-19&limit=100", "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5ultra_96:qwen38_27b", "rigId": "macstudio_m5ultra_96", "hardware": "Mac Studio M5 Ultra 96 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 96, "readingTps": 1075, "writingTps": 91, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": "2026-08-23", "seoEligible": true, "sourceUrls": [ "https://www.apple.com/shop/buy-mac/mac-studio/m5-ultra-chip-30-core-cpu-64-core-gpu-96gb-memory-1tb-storage" ], "estimateSource": "Conservative midpoint of bandwidth-scaled M3 Ultra and M5 Max Qwen3.8 source medians; not a Mac Studio M5 Ultra measurement", "estimateAnchors": [ { "factor": 1.9543973941368078, "rig": "m5max", "sourceBandwidth": 614, "targetBandwidth": 1200 }, { "factor": 1.5, "rig": "m3ultra", "sourceBandwidth": 800, "targetBandwidth": 1200 } ], "projectionMethod": null }, { "recordKey": "rtx4080super:qwen38_27b", "rigId": "rtx4080super", "hardware": "RTX 4080 Super", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 16, "readingTps": 1750, "writingTps": 43.6, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-08-18", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual5060:qwen38_27b", "rigId": "dual5060", "hardware": "2× RTX 5060 Ti", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 32, "readingTps": 723, "writingTps": 65.8, "readingStatus": "measured", "writingStatus": "measured", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 8, "serviceObservationCount": 0, "asOf": "2026-08-24", "seoEligible": true, "sourceUrls": [ "https://github.com/sudoingX/qwen38-mtp" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rx7900xtx:qwen38_27b", "rigId": "rx7900xtx", "hardware": "AMD Radeon RX 7900 XTX", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 24, "readingTps": 877.84, "writingTps": 47.195, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-08-17", "seoEligible": true, "sourceUrls": [ "https://github.com/sudoingX/qwen38-mtp", "https://smeltcore.com/recipes/qwen3-8-27b-on-rx-7900-xtx-128k-context-vision-chat-on-rocm-with-llama-cpp/" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx_a4000:qwen38_27b", "rigId": "rtx_a4000", "hardware": "RTX A4000", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 16, "readingTps": 925.925925925926, "writingTps": 28.3, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-08-18", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx3090:qwen38_27b", "rigId": "rtx3090", "hardware": "RTX 3090", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 24, "readingTps": 705, "writingTps": 53.15, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 10, "serviceObservationCount": 0, "asOf": "2026-08-23", "seoEligible": true, "sourceUrls": [ "https://jonidimo.github.io/qwen38-3090-benchmark/benchmark.html", "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24", "https://github.com/sudoingX/qwen38-mtp" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual4080:qwen38_27b", "rigId": "dual4080", "hardware": "2× RTX 4080", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 32, "readingTps": 1700, "writingTps": 58, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": null, "seoEligible": false, "sourceUrls": [], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dgx:qwen38_27b", "rigId": "dgx", "hardware": "DGX Spark", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 128, "readingTps": 1225.5, "writingTps": 44.0325, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 5, "serviceObservationCount": 0, "asOf": "2026-08-24", "seoEligible": true, "sourceUrls": [ "https://github.com/hasso5703/dgx-spark-qwen38", "https://github.com/0xBakeer/Qwen3.8-27B-4-bit-on-a-single-DGX-Spark/blob/main/RESULTS.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5max_128:qwen38_27b", "rigId": "macstudio_m5max_128", "hardware": "Mac Studio M5 Max 128 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 128, "readingTps": 745, "writingTps": 59.2, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": "2026-08-15", "seoEligible": true, "sourceUrls": [ "https://www.apple.com/shop/buy-mac/mac-studio/m5-max-chip-18-core-cpu-40-core-gpu-128gb-memory-1tb-storage" ], "estimateSource": "Projection anchored to the reported M5 Max Qwen3.8 source median; no Mac Studio measurement yet", "estimateAnchors": [ { "factor": 1, "rig": "m5max", "sourceBandwidth": 614, "targetBandwidth": 614 } ], "projectionMethod": null }, { "recordKey": "rtx5060:qwen38_27b", "rigId": "rtx5060", "hardware": "RTX 5060 Ti", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 16, "readingTps": 779.5100222717149, "writingTps": 27.4, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-08-18", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/Zynerji/Qwen3.8-27B-PristinelyUncensored-HOMEUSER-16-24" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "arcpro_b70:qwen38_27b", "rigId": "arcpro_b70", "hardware": "Intel Arc Pro B70", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 32, "readingTps": 424.8, "writingTps": 46.25, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 5, "serviceObservationCount": 0, "asOf": "2026-08-25", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/SergiioB/Qwen3.8-27B-GPTQ-Int4-sym-G128-MTP-BF16", "https://jonathanmann.tech/blog/qwen38-intel-arc-b70/", "https://github.com/vllm-project/vllm-xpu-kernels/issues/510", "https://github.com/SergiioB/intel-arc-pro-b70-inference-cookbook/issues/6" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m5max:qwen38_27b", "rigId": "m5max", "hardware": "MacBook Pro M5 Max 128 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 128, "readingTps": 745, "writingTps": 59.2, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-08-15", "seoEligible": true, "sourceUrls": [ "https://github.com/bluehawana/ollama-deepseek-v4-flash-iq2_m-mac-m5max" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m4max:qwen38_27b", "rigId": "m4max", "hardware": "Mac Studio M4 Max 64 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 64, "readingTps": 274, "writingTps": 53.3, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": null, "seoEligible": false, "sourceUrls": [], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtxpro6000:qwen38_27b", "rigId": "rtxpro6000", "hardware": "RTX PRO 6000 Blackwell", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 96, "readingTps": 7756, "writingTps": 109.25, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 4, "serviceObservationCount": 0, "asOf": "2026-08-25", "seoEligible": true, "sourceUrls": [ "https://www.nvidia.com/en-us/products/workstations/professional-desktop-gpus/rtx-pro-6000/", "https://github.com/sgl-project/sglang/issues/36118", "https://huggingface.co/lribeiro/Qwen3.8-27B-nvfp4-v6", "https://huggingface.co/pearsonkyle/Qwen3.8-27B-GPTQ-W4A16", "https://huggingface.co/Zynerji/Ektome-Qwen3.8-27B-PristinelyUncensored-GPTQ-MTP" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "r9700:qwen38_27b", "rigId": "r9700", "hardware": "AMD Radeon AI PRO R9700", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 32, "readingTps": 563.8, "writingTps": 35.15, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 8, "serviceObservationCount": 0, "asOf": "2026-08-18", "seoEligible": true, "sourceUrls": [ "https://www.amd.com/en/blogs/2026/run-qwen-3-8-27b-on-amd-ryzen-ai-max-and-radeon-graphics-cards-day-0.html", "https://github.com/sudoingX/qwen38-mtp", "https://github.com/rahlquist/qwen38-r9700" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5ultra_256:qwen38_27b", "rigId": "macstudio_m5ultra_256", "hardware": "Mac Studio M5 Ultra 256 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 256, "readingTps": 1075, "writingTps": 91, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": "2026-08-23", "seoEligible": true, "sourceUrls": [ "https://www.apple.com/shop/buy-mac/mac-studio/m5-ultra-chip-36-core-cpu-80-core-gpu-256gb-memory-1tb-storage" ], "estimateSource": "Conservative midpoint of bandwidth-scaled M3 Ultra and M5 Max Qwen3.8 source medians; not a Mac Studio M5 Ultra measurement", "estimateAnchors": [ { "factor": 1.9543973941368078, "rig": "m5max", "sourceBandwidth": 614, "targetBandwidth": 1200 }, { "factor": 1.5, "rig": "m3ultra", "sourceBandwidth": 800, "targetBandwidth": 1200 } ], "projectionMethod": null }, { "recordKey": "ryzen_ai_max_395_128:qwen38_27b", "rigId": "ryzen_ai_max_395_128", "hardware": "Ryzen AI Max+ 395 128 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 128, "readingTps": 300.03, "writingTps": 24.189999999999998, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-08-15", "seoEligible": true, "sourceUrls": [ "https://www.gmktec.com/products/amd-ryzen%E2%84%A2-ai-max-395-evo-x2-ai-mini-pc?variant=48746069262490", "https://www.amd.com/en/blogs/2026/run-qwen-3-8-27b-on-amd-ryzen-ai-max-and-radeon-graphics-cards-day-0.html", "https://www.localmaxxing.com/en/runs/cmstq9z3803q3ms01f0r9qsko" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macmini_m5pro_48:qwen38_27b", "rigId": "macmini_m5pro_48", "hardware": "Mac mini M5 Pro 48 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 48, "readingTps": 209.40000000000003, "writingTps": 14, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-08-20", "seoEligible": true, "sourceUrls": [ "https://omlx.ai/benchmarks/performance/tq3pj3dm", "https://omlx.ai/benchmarks/performance/pqocjn7h" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_v100_32gb:qwen38_27b", "rigId": "dual_v100_32gb", "hardware": "2× Tesla V100 32GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 64, "readingTps": 282.1, "writingTps": 38.3, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-08-26", "seoEligible": true, "sourceUrls": [ "https://www.localmaxxing.com/api/speed-tests?username=itgoatee&date=2026-08-26&limit=100" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m3ultra:qwen38_27b", "rigId": "m3ultra", "hardware": "Mac Studio M3 Ultra 512 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 512, "readingTps": 462.5, "writingTps": 44.25, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-08-23", "seoEligible": true, "sourceUrls": [ "https://omlx.ai/benchmarks/performance/5x3mzt9b", "https://github.com/sudoingX/qwen38-mtp" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m4max128:qwen38_27b", "rigId": "m4max128", "hardware": "Mac Studio M4 Max 128 GB", "modelId": "qwen38_27b", "model": "Qwen3.8 27B", "variantId": "qwen38_27b_c1_personal", "memoryGB": 128, "readingTps": null, "writingTps": 56.75, "readingStatus": "pending", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 7, "serviceObservationCount": 0, "asOf": "2026-08-21", "seoEligible": true, "sourceUrls": [ "https://github.com/Weschera/Qwen3.8-27B-oMLX-MTP-Mac" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5ultra_96:qwen38_flash_next", "rigId": "macstudio_m5ultra_96", "hardware": "Mac Studio M5 Ultra 96 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 96, "readingTps": 4083, "writingTps": 139.60000000000002, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-10-01", "seoEligible": true, "sourceUrls": [ "https://www.reddit.com/r/LocalLLM/comments/1wvki91/m5_mac_studio_ultra_96b_testing_qwen38_flash_next/" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx4090:qwen38_flash_next", "rigId": "rtx4090", "hardware": "RTX 4090", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 24, "readingTps": 4000, "writingTps": 109.30000000000001, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 4, "serviceObservationCount": 0, "asOf": "2026-10-05", "seoEligible": true, "sourceUrls": [ "https://github.com/FlashML-org/FreeToken/pull/257", "https://github.com/Niko1221/Strata/issues/675", "https://github.com/AnnoyingTechnology/nvidia-4090-flash-next-inference", "https://blog.ferstar.org/en/posts/qwen38-flash-next-strata-4090/" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx5080:qwen38_flash_next", "rigId": "rtx5080", "hardware": "RTX 5080", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 16, "readingTps": 3591.4, "writingTps": 80.05, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 10, "serviceObservationCount": 0, "asOf": "2026-10-09", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/500#issuecomment-5962920345", "https://github.com/Niko1221/Strata/issues/765#issuecomment-5981600251", "https://github.com/Niko1221/Strata/pull/1243" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "r9700:qwen38_flash_next", "rigId": "r9700", "hardware": "AMD Radeon AI PRO R9700", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 32, "readingTps": 2197.4, "writingTps": 80.8, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 6, "serviceObservationCount": 0, "asOf": "2026-10-12", "seoEligible": true, "sourceUrls": [ "https://github.com/zihaomu/Strata/blob/5562193c68fc5c88bff0710ebd28823b710a4298/bench/results/2026-10-08-community-r9700-linux/README.md", "https://github.com/Niko1221/Strata/pull/2047", "https://github.com/Niko1221/Strata/blob/7695859aa13afb248bdb4d36f9b2e95707db079d/docs/AMD_HIP.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx3090:qwen38_flash_next", "rigId": "rtx3090", "hardware": "RTX 3090", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 24, "readingTps": 2094, "writingTps": 86.45, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-10-05", "seoEligible": true, "sourceUrls": [ "https://github.com/FlashML-org/FreeToken/pull/232", "https://www.reddit.com/r/LocalLLaMA/comments/1wwamjw/strata_rtx_3090_128_ram_qwen_38_flash_next/", "https://github.com/Niko1221/Strata/pull/110" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx5090:qwen38_flash_next", "rigId": "rtx5090", "hardware": "RTX 5090", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 32, "readingTps": 6256.6, "writingTps": 173.775, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 5, "serviceObservationCount": 0, "asOf": "2026-10-10", "seoEligible": true, "sourceUrls": [ "https://github.com/FlashML-org/FreeToken/pull/257", "https://github.com/Niko1221/Strata/pull/811", "https://github.com/Niko1221/Strata/issues/1791" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx5070:qwen38_flash_next", "rigId": "rtx5070", "hardware": "RTX 5070 Ti", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 16, "readingTps": 2498.45, "writingTps": 71.1, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 16, "serviceObservationCount": 0, "asOf": "2026-10-05", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/issues/791", "https://github.com/Zauberio/Strata/blob/31fc22854bc6aad8ebee3d4d2e87d59efd1fe9f5/bench/results/2026-10-04-community-rtx-5070ti-windows/README.md", "https://llamaperf.com/gpu/rtx-5070-ti", "https://github.com/obiscr/Strata/blob/36ddda197895b09895b5bcbb956813b35fce219b/bench/results/2026-10-06-community-rtx5070ti-5900x/README.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx4070ti_super:qwen38_flash_next", "rigId": "rtx4070ti_super", "hardware": "RTX 4070 Ti SUPER 16GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 16, "readingTps": 1700, "writingTps": 72, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-10-06", "seoEligible": true, "sourceUrls": [ "https://www.reddit.com/r/LocalLLM/comments/1wy9uc7/strata_tested_on_rtx_4070_ti_super_ryzen_9/" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual5070:qwen38_flash_next", "rigId": "dual5070", "hardware": "2× RTX 5070 Ti", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 32, "readingTps": 3030, "writingTps": 96, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-10-15", "seoEligible": true, "sourceUrls": [ "https://llamaperf.com/gpu/rtx-5070-ti" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx3090ti:qwen38_flash_next", "rigId": "rtx3090ti", "hardware": "RTX 3090 Ti", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 24, "readingTps": 2255.1, "writingTps": 93.1, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": "2026-10-05", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/110" ], "estimateSource": "RTX 3090 Ti Flash Next Strata estimate scaled from the RTX 3090 median by the slower of the memory-bandwidth ratio (1,008/936 = 1.077) and the dense FP32 ratio (40/35.58 = 1.124): 2,094 x 1.077 = 2,255 tok/s reading and 86.45 x 1.077 = 93.1 tok/s writing. The only RTX 3090 Ti measurements (Strata PR #674, 1,239-1,304 tok/s reading and 67.1-98.8 tok/s writing) were taken on a dual-socket NUMA server that the curator ruled unrepresentative of a desktop card; they are excluded from the headline and retained in the private evidence store.", "estimateAnchors": [ { "factor": 1.0769230769230769, "limiting": [ "bandwidth" ], "rig": "rtx3090", "sourceBandwidth": 936, "sourceCondition": "RTX 3090 Strata IQ3_S median over both ingested arms: 1,361 / 45 tok/s on engine 0.1.13 (Strata PR #110) and 2,827 / 127.9 tok/s on engine 0.1.39 (ModelDock operator machine, Ryzen 9 9950X, 128 GB DDR5).", "sourceDate": "5 Oct 2026", "sourceDecode": 86.45, "sourcePrefill": 2094, "sourceUrl": "https://github.com/Niko1221/Strata/pull/110", "targetBandwidth": 1008 } ], "projectionMethod": "conservative-hardware-ratio" }, { "recordKey": "rx7900xtx:qwen38_flash_next", "rigId": "rx7900xtx", "hardware": "AMD Radeon RX 7900 XTX", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 24, "readingTps": 1529.2, "writingTps": 71.2, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 10, "serviceObservationCount": 0, "asOf": "2026-10-13", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/745", "https://github.com/Niko1221/Strata/pull/1173", "https://github.com/Niko1221/Strata/issues/1831", "https://github.com/Niko1221/Strata/issues/2274" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_v100_32gb:qwen38_flash_next", "rigId": "dual_v100_32gb", "hardware": "2× Tesla V100 32GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 64, "readingTps": 2355.3427926568047, "writingTps": 78.8, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-10-08", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/1508" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rx9070xt_16gb:qwen38_flash_next", "rigId": "rx9070xt_16gb", "hardware": "AMD Radeon RX 9070 XT 16GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 16, "readingTps": 1660, "writingTps": 54.9, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 6, "serviceObservationCount": 0, "asOf": "2026-10-11", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/issues/1211", "https://github.com/Niko1221/Strata/pull/1998" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5ultra_256:qwen38_flash_next", "rigId": "macstudio_m5ultra_256", "hardware": "Mac Studio M5 Ultra 256 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 256, "readingTps": 3441.5, "writingTps": 155, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 13, "serviceObservationCount": 0, "asOf": "2026-10-15", "seoEligible": true, "sourceUrls": [ "https://github.com/sethforprivacy/Qwen3.8-Flash-Next-M5-Ultra-oMLX", "https://www.linkedin.com/posts/ivanfioravanti_m5-ultra-vs-2-dgx-sparks-the-number-youre-activity-7511826876328419329-n8tG", "https://github.com/ddalcu/mlx-serve/pull/685#issuecomment-6016456323", "https://github.com/ddalcu/mlx-serve/blob/main/benchmarks.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual5060:qwen38_flash_next", "rigId": "dual5060", "hardware": "2× RTX 5060 Ti", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 32, "readingTps": 1602, "writingTps": 67.6, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 11, "serviceObservationCount": 0, "asOf": "2026-10-05", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/483", "https://github.com/Niko1221/Strata/issues/791" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx5070_12gb:qwen38_flash_next", "rigId": "rtx5070_12gb", "hardware": "RTX 5070 12 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 12, "readingTps": 1443.21, "writingTps": 48.31, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 5, "serviceObservationCount": 0, "asOf": "2026-09-29", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/blob/6f32ec070f23ced9f50e704d854d775da52591ab/bench/results/2026-09-29-speed-0126/README.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_rtx3090_24gb:qwen38_flash_next", "rigId": "dual_rtx3090_24gb", "hardware": "2x RTX 3090 24 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 48, "readingTps": 2137.55, "writingTps": 111.3, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 10, "serviceObservationCount": 0, "asOf": "2026-10-06", "seoEligible": true, "sourceUrls": [ "https://github.com/TreyThomasCodes/strata-local/blob/main/SETTINGS.md", "https://github.com/Niko1221/Strata", "https://github.com/Niko1221/Strata/pull/110", "https://github.com/q8atnight/strata-dual-3090/blob/17553ca1c582a196e21a489c3643f4027eea3d15/dual3090/README.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5max_128:qwen38_flash_next", "rigId": "macstudio_m5max_128", "hardware": "Mac Studio M5 Max 128 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 128, "readingTps": 2146.5, "writingTps": 65.24, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-10-04", "seoEligible": true, "sourceUrls": [ "https://lcz.me/topic/1997/11" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "ryzen_ai_max_395_128:qwen38_flash_next", "rigId": "ryzen_ai_max_395_128", "hardware": "Ryzen AI Max+ 395 128 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 128, "readingTps": 1242.707970267098, "writingTps": 47.48976023935716, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 10, "serviceObservationCount": 0, "asOf": "2026-10-09", "seoEligible": true, "sourceUrls": [ "https://www.soothill.io/blog/2026/08/27/qwen38-flash-next-rocm-vulkan-strix-halo/", "https://github.com/Niko1221/Strata/releases/tag/v0.1.40", "https://github.com/Niko1221/Strata/pull/917" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_r9700_32gb:qwen38_flash_next", "rigId": "dual_r9700_32gb", "hardware": "2× AMD Radeon AI PRO R9700 32 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 64, "readingTps": 4244.5, "writingTps": 79.15, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 4, "serviceObservationCount": 0, "asOf": "2026-10-06", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/issues/1008" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "arcpro_b70:qwen38_flash_next", "rigId": "arcpro_b70", "hardware": "Intel Arc Pro B70", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 32, "readingTps": 782.7875, "writingTps": 51.435, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 11, "serviceObservationCount": 0, "asOf": "2026-10-07", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/423#issuecomment-5980039356", "https://github.com/Niko1221/Strata/blob/b2c5b221317c9db365cdbfff290edb4aa5557911/docs/INTEL.md" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m5max:qwen38_flash_next", "rigId": "m5max", "hardware": "MacBook Pro M5 Max 128 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 128, "readingTps": 1245.5, "writingTps": 77.6, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-10-13", "seoEligible": true, "sourceUrls": [ "https://blog.brainosi.com/en/2026-09-01-qwen38-flash-next-m5max-prefill/", "https://x.com/Beamsters1/status/2096397815010439225", "https://github.com/Niko1221/Strata" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dgx:qwen38_flash_next", "rigId": "dgx", "hardware": "DGX Spark", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 128, "readingTps": 2228.5, "writingTps": 48.7, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 7, "serviceObservationCount": 0, "asOf": "2026-09-06", "seoEligible": true, "sourceUrls": [ "https://github.com/0xBakeer/qwen38-flash-next-spark", "https://huggingface.co/Baekpica/Qwen3.8-Flash-Next-Mixed-Quant-SSD-PLE-GGUF", "https://github.com/MiaAI-Lab/Qwen3.8-Flash-Next-Single-DGX-Spark" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_titan_rtx_24gb:qwen38_flash_next", "rigId": "dual_titan_rtx_24gb", "hardware": "2x TITAN RTX 24 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 48, "readingTps": 1713.35, "writingTps": 66.1, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 4, "serviceObservationCount": 0, "asOf": "2026-10-08", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/389", "https://github.com/Niko1221/Strata" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "arcpro_b60:qwen38_flash_next", "rigId": "arcpro_b60", "hardware": "Intel Arc Pro B60 24GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 24, "readingTps": 1034, "writingTps": 35.5, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 0, "asOf": "2026-10-14", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/issues/2339" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_rtx3060_12gb:qwen38_flash_next", "rigId": "dual_rtx3060_12gb", "hardware": "2x RTX 3060 12 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 24, "readingTps": 1239.8, "writingTps": 41.25, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-10-04", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/721" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtxpro6000:qwen38_flash_next", "rigId": "rtxpro6000", "hardware": "RTX PRO 6000 Blackwell", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 96, "readingTps": 8987.925, "writingTps": 171.09, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 3, "serviceObservationCount": 1, "asOf": "2026-08-30", "seoEligible": true, "sourceUrls": [ "https://github.com/jpezzulli/sglang-rtxpro6000/blob/pennyroyal-main-sm120-final/RESULTS.md#flash-next-final-campaign" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rx6900xt_16gb:qwen38_flash_next", "rigId": "rx6900xt_16gb", "hardware": "AMD Radeon RX 6900 XT 16GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 16, "readingTps": 465.95, "writingTps": 47.599999999999994, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-10-06", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/1270" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "dual_rx6900xt_16gb:qwen38_flash_next", "rigId": "dual_rx6900xt_16gb", "hardware": "2x AMD Radeon RX 6900 XT 16GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 32, "readingTps": 772.3, "writingTps": 63, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 2, "serviceObservationCount": 0, "asOf": "2026-10-06", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/1270" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx3060_12gb:qwen38_flash_next", "rigId": "rtx3060_12gb", "hardware": "RTX 3060 12 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 12, "readingTps": 819.9543580093687, "writingTps": 20.2, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 6, "serviceObservationCount": 0, "asOf": "2026-10-09", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/pull/1725" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "macstudio_m5max_64:qwen38_flash_next", "rigId": "macstudio_m5max_64", "hardware": "Mac Studio M5 Max 64 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 64, "readingTps": 517.9, "writingTps": 36, "readingStatus": "estimate", "writingStatus": "estimate", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Estimate", "estimate": true, "partialEstimate": false, "sourceBacked": false, "sourceObservationCount": 0, "serviceObservationCount": 0, "asOf": "2026-08-28", "seoEligible": true, "sourceUrls": [ "https://www.apple.com/shop/buy-mac/mac-studio/m5-max-chip-18-core-cpu-40-core-gpu-64gb-memory-1tb-storage" ], "estimateSource": "Mac Studio M5 Max Flash Next projection from the source-backed M5 Max 64GB result; capped by the slowest CPU-core, GPU-core, or unified-memory-bandwidth ratio. Not an exact-machine measurement.", "estimateAnchors": [ { "factor": 1, "limiting": [ "cpu", "gpu", "bandwidth" ], "rig": "apple_m5max_macbook_64_atomic", "sourceBandwidth": 614, "sourceCpuCores": 18, "sourceGpuCores": 40, "targetBandwidth": 614, "targetCpuCores": 18, "targetGpuCores": 40 } ], "projectionMethod": "conservative-hardware-ratio" }, { "recordKey": "macmini_m5pro_64:qwen38_flash_next", "rigId": "macmini_m5pro_64", "hardware": "Mac mini M5 Pro 64 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 64, "readingTps": 430, "writingTps": 19.23, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 4, "serviceObservationCount": 0, "asOf": "2026-10-10", "seoEligible": true, "sourceUrls": [ "https://github.com/skeggsguy/Flash-next-ssd" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "apple_m5max_macbook_64_atomic:qwen38_flash_next", "rigId": "apple_m5max_macbook_64_atomic", "hardware": "MacBook Pro M5 Max 64 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 64, "readingTps": 517.9, "writingTps": 36, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-08-28", "seoEligible": true, "sourceUrls": [ "https://huggingface.co/AtomicChat/Qwen3.8-Flash-Next-GGUF" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m4max128:qwen38_flash_next", "rigId": "m4max128", "hardware": "Mac Studio M4 Max 128 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 128, "readingTps": 446.819955349363, "writingTps": 32.46163434301431, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-08-31", "seoEligible": true, "sourceUrls": [ "https://github.com/snagnever/macstudio-local-llm/blob/main/bench/qwen38-flash-next/results/refresh-flashnext-128k-v064-ssdple.jsonl" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "m3ultra:qwen38_flash_next", "rigId": "m3ultra", "hardware": "Mac Studio M3 Ultra 512 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 512, "readingTps": 479.6019334885481, "writingTps": 18.101478141358502, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-09-07", "seoEligible": true, "sourceUrls": [ "https://github.com/zuyezheng/local-llm-bench/blob/main/analysis/samples_tidy.csv" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx4070:qwen38_flash_next", "rigId": "rtx4070", "hardware": "RTX 4070 12 GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 12, "readingTps": 1376, "writingTps": 36.9, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 1, "serviceObservationCount": 0, "asOf": "2026-10-04", "seoEligible": true, "sourceUrls": [ "https://note.com/fika_semla/n/n1801b1ee1dd6" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null }, { "recordKey": "rtx4080_16gb:qwen38_flash_next", "rigId": "rtx4080_16gb", "hardware": "RTX 4080 16GB", "modelId": "qwen38_flash_next", "model": "Qwen3.8 Flash Next", "variantId": "qwen38_flash_next_c1_personal", "memoryGB": 16, "readingTps": 1996.3, "writingTps": 66, "readingStatus": "reported", "writingStatus": "reported", "readingEstimateLow": null, "readingEstimateHigh": null, "writingEstimateLow": null, "writingEstimateHigh": null, "evidenceStatus": "Median", "estimate": false, "partialEstimate": false, "sourceBacked": true, "sourceObservationCount": 4, "serviceObservationCount": 0, "asOf": "2026-10-12", "seoEligible": true, "sourceUrls": [ "https://github.com/Niko1221/Strata/issues/2117" ], "estimateSource": null, "estimateAnchors": [], "projectionMethod": null } ] }