Model benchmarks
hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S local LLM performance
As of October 2026, hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S runs at up to 34.4 tok/s for local inference (best of 4 community benchmark runs across 2 GPUs).
Model size
26.9B
Peak speed
34.4 tok/s
Average speed
29.2 tok/s
Avg PP
326.3 tok/s
Min memory
17.6 GB
Max context
262,144 tokens
Avg output / run
60,020 tokens
Avg runtime / run
37m 44s
Avg quality
85.1
Benchmark runs
4
GPUs tested
2
Quality by task
Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 4 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.
| Task | P5 (low) | Avg | P95 (high) |
|---|---|---|---|
| Overall | 84.4 | 85.1 | 85.8 |
| Agent Workflow | 77.2 | 83.2 | 90.3 |
| Code Generation | 73.6 | 78.3 | 85.0 |
| Role Play & Narrative | 90.2 | 92.8 | 94.5 |
| Research & Analysis | 83.7 | 86.1 | 88.1 |
Performance by hardware and tool
Every hardware/tool/quantization combination hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S has been benchmarked on, ranked by peak token generation speed. Last updated October 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7900 XTX | Ollama | IQ3_S | 34.4 tok/s | 27.6 tok/s | 18.3 GB | 180,000 tokens | 85.3 | 3 |
| Apple M4 Pro | Splash | — | 34.0 tok/s | 34.0 tok/s | 17.6 GB | 262,144 tokens | 84.3 | 1 |
Benchmark runs
All 4 hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S good for coding?
- In our benchmarks, hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S scores 78.3/100 for coding. It runs at about 29.2 tok/s, so if you want more speed, Qwen3.8-27B-Turbo-Brilliance-50X-RI-GSQ-RCO-IQ3_XXS-mtp is faster (~37.5 tok/s) and still scores well for coding (83.8/100).
- Is hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S good for agentic (tool-using) tasks?
- In our benchmarks, hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S scores 83.2/100 for agentic workflows. It runs at about 29.2 tok/s, so if you want more speed, swift-iq2_xs is faster (~91.5 tok/s) and still scores well for agentic workflows (91.4/100).
- How fast is hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S for local inference?
- Across 4 community benchmark runs, hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S reaches up to 34.4 tok/s and averages 29.2 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S need?
- The leanest observed configuration used about 17.6 GB of memory (quantizations tested: IQ3_S).
- Which tools have been used to run hf.co/ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S?
- Benchmarks were submitted using Ollama, Splash. Results are community-contributed and updated as new runs arrive.