Model benchmarks
strata_swift local LLM performance
As of November 2026, strata_swift runs at up to 57.0 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
57.0 tok/s
Average speed
56.4 tok/s
Avg PP
409.6 tok/s
Min memory
97.6 GB
Max context
262,144 tokens
Avg output / run
43,804 tokens
Avg runtime / run
12m 18s
Avg quality
84.4
Benchmark runs
4
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 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 | 83.5 | 84.4 | 85.0 |
| Agent Workflow | 91.3 | 91.8 | 92.9 |
| Code Generation | 63.1 | 64.8 | 66.4 |
| Role Play & Narrative | 92.2 | 92.6 | 93.2 |
| Research & Analysis | 87.0 | 88.3 | 89.7 |
Performance by hardware and tool
Every hardware/tool/quantization combination strata_swift has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7900 XTX | OpenAI-compatible | — | 57.0 tok/s | 56.4 tok/s | 97.6 GB | 262,144 tokens | 84.4 | 4 |
Benchmark runs
All 4 strata_swift runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is strata_swift good for coding?
- In our benchmarks, strata_swift scores 64.8/100 for coding. It runs at about 56.4 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
- Is strata_swift good for agentic (tool-using) tasks?
- In our benchmarks, strata_swift scores 91.8/100 for agentic workflows. It runs at about 56.4 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
- How fast is strata_swift for local inference?
- Across 4 community benchmark runs, strata_swift reaches up to 57.0 tok/s and averages 56.4 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does strata_swift need?
- The leanest observed configuration used about 97.6 GB of memory.
- Which tools have been used to run strata_swift?
- Benchmarks were submitted using OpenAI-compatible. Results are community-contributed and updated as new runs arrive.