Better than 56% of all models
Context
N/A
Input $/1M
TBD
Output $/1M
TBD
Type
text-generation
License
Open Source
Benchmarks
7 tested
Data updated today
About
Microsoft text generation model. 744K downloads on HuggingFace.
Tested on 7 benchmarks with 27.6% average. Top scores: Chatbot Arena Elo — Overall (1127.2%), IFEval (54.8%), BBH (HuggingFace) (36.6%).
Capabilities
reasoning
13.1
#119 globally
math
16.4
#179 globally
knowledge
22.3
#189 globally
language
54.8
#100 globally
general
36.6
#22 globally
Benchmark Scores
Compare AllTested on 7 benchmarks · Ranked across 6 categories
Score Distribution (all 231 models)
0255075100
▲ You are here
reasoningCompare reasoning →
MUSR
13.1—HuggingFace MuSR (Multi-Step Reasoning). Tests multi-hop reasoning requiring chaining multiple facts together.
mathCompare math →
MATH Level 5
16.4—HuggingFace evaluation of MATH Level 5 problems. Competition math requiring advanced reasoning and proof construction.
knowledgeCompare knowledge →
MMLU-PRO
33.6—HuggingFace MMLU-Pro. Harder version of MMLU with 10 answer choices instead of 4 and more challenging questions.
GPQA
11.0—HuggingFace evaluation of GPQA (Graduate-Level Google-Proof Q&A). PhD-level science questions that cannot be easily searched.
vs Magnum v4 72B
vs Claude Sonnet 4.5 Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Links
Info
Research
Documentation
Community
Source Code
BenchGecko API
microsoft-phi-3-mini-4k-instruct
Specifications
- Typetext-generation
- ContextN/A
- ReleasedApr 2024
- LicenseOpen Source
- StatusActive
Available On
MicrosoftTBDLearn More
Share & Export
Frequently Asked Questions
Phi 3 Mini 4k Instruct is an open-source text-generation AI model by Microsoft, released in April 2024. It has an average benchmark score of 51.1.