Model benchmarks
gpt-oss-20b-Q4_K_M local LLM performance
As of September 2026, gpt-oss-20b-Q4_K_M runs at up to 142.0 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
20B
Peak speed
142.0 tok/s
Average speed
132.8 tok/s
Avg prefill
2090.4 tok/s
Min memory
9.8 GB
Max context
131,072 tokens
Avg output / run
5,348 tokens
Avg runtime / run
42s
Avg quality
64.9
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 | 62.0 | 64.9 | 67.7 |
| Agent Workflow | 74.6 | 79.4 | 84.1 |
| Code Generation | 53.0 | 57.8 | 62.5 |
| Role Play & Narrative | 58.3 | 59.2 | 60.0 |
| Research & Analysis | 61.8 | 63.1 | 64.5 |
Performance by hardware and tool
Every hardware/tool/quantization combination gpt-oss-20b-Q4_K_M has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 5060 | llama.cpp | Q4_K_M | 142.0 tok/s | 132.8 tok/s | 9.8 GB | 131,072 tokens | 64.9 | 2 |
Frequently asked questions
- Is gpt-oss-20b-Q4_K_M good for coding?
- In our benchmarks, gpt-oss-20b-Q4_K_M scores 57.8/100 for coding. It runs at about 132.8 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 gpt-oss-20b-Q4_K_M good for agentic (tool-using) tasks?
- In our benchmarks, gpt-oss-20b-Q4_K_M scores 79.4/100 for agentic workflows. It runs at about 132.8 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~165.0 tok/s) and still scores well for agentic workflows (89.3/100).
- How fast is gpt-oss-20b-Q4_K_M for local inference?
- Across 2 community benchmark runs, gpt-oss-20b-Q4_K_M reaches up to 142.0 tok/s and averages 132.8 tok/s, with the fastest results on NVIDIA GeForce RTX 5060.
- How much memory does gpt-oss-20b-Q4_K_M need?
- The leanest observed configuration used about 9.8 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run gpt-oss-20b-Q4_K_M?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.