Model benchmarks
SC117-Swift-IQ3_S local LLM performance
As of October 2026, SC117-Swift-IQ3_S runs at up to 114.4 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
114.4 tok/s
Average speed
111.5 tok/s
Avg PP
641.6 tok/s
Min memory
86.3 GB
Max context
262,144 tokens
Avg output / run
42,730 tokens
Avg runtime / run
6m 10s
Avg quality
79.5
Benchmark runs
2
GPUs tested
1
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 2 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 | 74.1 | 79.5 | 85.0 |
| Agent Workflow | 88.0 | 90.1 | 92.2 |
| Code Generation | 27.2 | 46.8 | 66.4 |
| Role Play & Narrative | 92.7 | 93.2 | 93.6 |
| Research & Analysis | 87.9 | 88.1 | 88.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination SC117-Swift-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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA RTX A6000 | llama.cpp | — | 114.4 tok/s | 111.5 tok/s | 86.3 GB | 262,144 tokens | 79.6 | 2 |
Benchmark runs
All 2 SC117-Swift-IQ3_S runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is SC117-Swift-IQ3_S good for coding?
- In our benchmarks, SC117-Swift-IQ3_S scores 46.8/100 for coding. It runs at about 111.5 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
- Is SC117-Swift-IQ3_S good for agentic (tool-using) tasks?
- In our benchmarks, SC117-Swift-IQ3_S scores 90.1/100 for agentic workflows. It runs at about 111.5 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~162.4 tok/s) and still scores well for agentic workflows (87.8/100).
- How fast is SC117-Swift-IQ3_S for local inference?
- Across 2 community benchmark runs, SC117-Swift-IQ3_S reaches up to 114.4 tok/s and averages 111.5 tok/s, with the fastest results on NVIDIA RTX A6000.
- How much memory does SC117-Swift-IQ3_S need?
- The leanest observed configuration used about 86.3 GB of memory.
- Which tools have been used to run SC117-Swift-IQ3_S?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.