Model benchmarks
default local LLM performance
As of November 2026, default runs at up to 68.1 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
68.1 tok/s
Average speed
56.2 tok/s
Avg PP
297.1 tok/s
Min memory
97.9 GB
Max context
524,288 tokens
Avg output / run
47,222 tokens
Avg runtime / run
14m 50s
Avg quality
77.5
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 | 62.7 | 77.5 | 85.9 |
| Agent Workflow | 83.7 | 89.5 | 93.0 |
| Code Generation | 17.9 | 52.4 | 76.1 |
| Role Play & Narrative | 83.8 | 89.8 | 93.2 |
| Research & Analysis | 56.5 | 78.3 | 89.1 |
Performance by hardware and tool
Every hardware/tool/quantization combination default 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 | — | 68.1 tok/s | 56.2 tok/s | 97.9 GB | 524,288 tokens | 77.5 | 4 |
Benchmark runs
All 4 default runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is default good for coding?
- In our benchmarks, default scores 52.4/100 for coding. It runs at about 56.2 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 default good for agentic (tool-using) tasks?
- In our benchmarks, default scores 89.6/100 for agentic workflows. It runs at about 56.2 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 default for local inference?
- Across 4 community benchmark runs, default reaches up to 68.1 tok/s and averages 56.2 tok/s, with the fastest results on AMD Radeon RX 7900 XTX.
- How much memory does default need?
- The leanest observed configuration used about 97.9 GB of memory.
- Which tools have been used to run default?
- Benchmarks were submitted using OpenAI-compatible. Results are community-contributed and updated as new runs arrive.