Configs › SuperQwen3.8-27B-abliterated on DGX Spark: 25.8 tok/s measured
SuperQwen3.8-27B-abliterated on DGX Spark: 25.8 tok/s measured
Measured SuperQwen3.8-27B-abliterated speeds: 25.8 tok/s on DGX Spark. 2 configs, each linked to its source, snapshot 2026-11-01.
Snapshot 2026-11-01 · 412 configs on AI boxes and 288 on GPU cards · 121 models. Every number is generated from the tracker snapshot and linked to where it was measured; nothing here is typed by hand.
Fastest measured single-stream run of SuperQwen3.8-27B-abliterated: 25.8 tok/s on DGX Spark (NVFP4, vllm, speculative decode, spec-k5), snapshot 2026-11-01.
Fastest measured run on each box
| Hardware | Decode tok/s | Quant | Backend | Context | Trust | Configs on this box | Source |
|---|---|---|---|---|---|---|---|
| DGX Spark | 25.8 | NVFP4 | vllm (speculative decode, spec-k5) | 262043 | Extracted, source-linked | 2 | source |
Questions and answers
- How fast is SuperQwen3.8-27B-abliterated on local hardware?
- The fastest measured single-stream run is 25.8 tok/s on DGX Spark (NVFP4, vllm, speculative decode, spec-k5). TokenMark tracks 2 single-stream configs for SuperQwen3.8-27B-abliterated on 1 box.
- How many measured configs does TokenMark have for SuperQwen3.8-27B-abliterated?
- 2 single-stream configs on 1 box, newest source check 2026-11-01. Every row links to where it was measured.
All measured configs for SuperQwen3.8-27B-abliterated
| Model | Hardware | Quant | Backend | Decode | Context | Trust | Source |
|---|---|---|---|---|---|---|---|
| SuperQwen3.8-27B-abliterated | DGX Spark | NVFP4 | vllm (speculative decode, spec-k5) | 25.8 tok/s | 262043 | Extracted, source-linked | source |
| SuperQwen3.8-27B-abliterated | DGX Spark | NVFP4 | vllm | 17.3 tok/s | Structured table | source |
Last updated: snapshot 2026-11-01; newest source check 2026-11-01. Rankings use single-stream runs only; aggregate and multi-GPU rows are kept off them.
More: all configs · hardware · news · contributors · methodology · llms.txt · llms-full.txt · API (OpenAPI) · JSON snapshot. Built by Altronis, private on-prem AI, Singapore.