Phi-4-mini-instruct: benchmark scores, price and ranking
Phi-4-mini-instruct from microsoft ranks #721 of 743 LLMs on the LLMs Tiger aggregate benchmark score (41.8), based on 3 benchmark results captured up to 2026-10-13.
- Organisation
- microsoft
- Type
- Open-weight / local model
- Open weights
- Yes
- Released
- 2025-02-27 (source: epoch)
- Cheapest API price
- $0.075 in / $0.300 out per million tokens (via azure_ai, observed 2026-10-13T21:47Z)
- Aggregate rank
- #721 of 743 (score 41.8)
Relative to other models, Phi-4-mini-instruct is strongest on MMLU-Pro (#71 of 136) and weakest on SciCode (#106 of 106).
Across the 3 benchmarks where at least ten models are scored, Phi-4-mini-instruct is in the top 10% on 0 and in the top half on 0.
Its cheapest listed input price, $0.075 per million tokens, is lower than that of 87% of the 427 models with a listed price.
Phi-4-mini-instruct benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| MMLU-Pro | 52.8% | #71 of 136 | HuggingFace MMLU-Pro leaderboard | 2026-10-13 |
| CritPt | 0.0% | #110 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 10.8% | #106 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
About Phi-4-mini-instruct
Phi-4-mini-instruct is an open-weight model that can run on your own hardware from microsoft. It was released on 2025-02-27, according to Epoch AI. The cheapest API listing we track charges $0.075 per million input tokens and $0.300 per million output tokens, through azure_ai. We hold 3 benchmark results for it across 2 areas, and none is in the top tenth of its board.
Where Phi-4-mini-instruct stands
On our aggregate score, Phi-4-mini-instruct is #721 of 743 ranked models, in the lower half. Its score is 41.8; the scale is centred on 50, so it sits 8.2 points below the typical model. That beats 3% of the ranked field. Among the 4 ranked microsoft models it is #4. The next one up is Phi-3-mini-128k-instruct at 46.3.
Phi-4-mini-instruct, benchmark by benchmark
Phi-4-mini-instruct lands below the middle on MMLU-Pro: 52.8%, #71 of 136, with 51% of the field scoring higher. Darwin-180B-RSI leads at 88.1%; the closest result above is Mistral-Small-24B-Base-2501 at 54.4%. It covers 57-subject knowledge + reasoning. The board median is 55.3%, so Phi-4-mini-instruct is 2.5 points below it. Source: HuggingFace MMLU-Pro leaderboard, captured 2026-10-13.
CritPt is a weak spot. Phi-4-mini-instruct scores 0.0%, #110 of 115, behind 95% of the models measured. The leader, GPT-5.6 Sol (max), reaches 32.3%. It covers unpublished research-level physics problems. The board median is 2.9%, so Phi-4-mini-instruct is 2.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
SciCode is a weak spot. Phi-4-mini-instruct scores 10.8%, #106 of 106, behind 99% of the models measured. The leader, Claude Opus 5.5 (max), reaches 66.9%. It covers scientific research code. The board median is 44.3%, so Phi-4-mini-instruct is 33.6 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Reasoning, maths and science: weak, ahead of 26% of the field on average across 2 benchmarks. Its best board here is MMLU-Pro (#71 of 136) and its weakest is CritPt (#110 of 115).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so Phi-4-mini-instruct costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on Llama 3.2 1B ($0.100) and Mistral-7B-Instruct-v0.3 ($0.100).
Reading these numbers
The 3 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (2), HuggingFace MMLU-Pro leaderboard (1). No official leaderboard supplied any of them, so treat them as third-party measurements. With 3 results the aggregate rank can move a lot when one new score arrives.
Models ranked nearby
- #718 Mistral-7B-Instruct-v0.3 (41.9)
- #719 mistral-small-2503 (41.9)
- #720 claude-3-sonnet-20240229 (41.9)
- #722 granite-4.0-350m (41.6)
- #723 gemini-1.0-pro-001 (41.5)
- #724 Claude 3.5 Haiku (Oct 2024) (41.3)
More from microsoft
- Phi-3-medium-4k-instruct (#321, 48.8)
- Phi-3.5-mini-instruct (#507, 46.5)
- Phi-3-mini-128k-instruct (#527, 46.3)
Compare it on the live leaderboard → · How the aggregate score works