Qwen3.8 27Bqwen38_27b_c1_personal
1,225.5 tok/s
44 tok/s
Source-backed
ModelDock / hardware record
DGX Spark local AI benchmarks
Compare the published model-specific reading / prefill and single-stream writing / decode results for DGX Spark.
Memory128 GBPublished model records: 2
Model-independent specification
Durable hardware facts.
Configured capacity and peak physical bandwidth feed the Hardware ranking. Model throughput remains in the hardware × model records below.
Memory128 GBUnified
Peak bandwidth273 GB/sUnified memory
Form factorAI box1 ×
Specification statusOfficial specOpen public specification ↗
Available model records
DGX Spark performance by model
ModelReading / prefillWriting / decodeEvidenceAs of
Qwen3.8 Flash Nextqwen38_flash_next_c1_personal
2,228.5 tok/s
48.7 tok/s
Source-backed
Reading / prefill
Prefill · tok/s
Writing / decode
Decode · single stream · tok/s
Evidence
Each row is one hardware × model record. Source observations stay on Evidence; projections link to their calibration details on Data.