Model benchmarks
phi-4-mini-instruct local LLM performance
As of November 2026, phi-4-mini-instruct runs at up to 86.4 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).
Model size
Unknown
Peak speed
86.4 tok/s
Average speed
85.6 tok/s
Avg PP
4169.3 tok/s
Min memory
6.0 GB
Max context
65,536 tokens
Avg output / run
4,171 tokens
Avg runtime / run
1m 8s
Avg quality
28.5
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 | 25.9 | 28.5 | 31.9 |
| Agent Workflow | 12.6 | 28.1 | 39.4 |
| Code Generation | 0.0 | 5.3 | 11.9 |
| Role Play & Narrative | 42.7 | 47.6 | 52.9 |
| Research & Analysis | 27.0 | 33.2 | 39.6 |
Performance by hardware and tool
Every hardware/tool/quantization combination phi-4-mini-instruct 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 3060 Laptop GPU | Bionic | — | 86.4 tok/s | 85.6 tok/s | 6.0 GB | 65,536 tokens | 28.5 | 4 |
Benchmark runs
All 4 phi-4-mini-instruct runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is phi-4-mini-instruct good for coding?
- In our benchmarks, phi-4-mini-instruct scores 5.3/100 for coding. It runs at about 85.6 tok/s, so if you want more speed, ornith-1.5-35b-a3b-apex-mtp-i-quality is faster (~105.2 tok/s) and still scores well for coding (83.6/100).
- Is phi-4-mini-instruct good for agentic (tool-using) tasks?
- In our benchmarks, phi-4-mini-instruct scores 28.1/100 for agentic workflows. It runs at about 85.6 tok/s, so if you want more speed, ornith-1.5-35b-a3b-apex-mtp-quality is faster (~103.2 tok/s) and still scores well for agentic workflows (88.5/100).
- How fast is phi-4-mini-instruct for local inference?
- Across 4 community benchmark runs, phi-4-mini-instruct reaches up to 86.4 tok/s and averages 85.6 tok/s, with the fastest results on NVIDIA GeForce RTX 3060 Laptop GPU.
- How much memory does phi-4-mini-instruct need?
- The leanest observed configuration used about 6.0 GB of memory.
- Which tools have been used to run phi-4-mini-instruct?
- Benchmarks were submitted using Bionic. Results are community-contributed and updated as new runs arrive.