reasoning
Phi-4-mini-flash-reasoning
Microsoft 3.85B reasoning model whose card names lightweight deployment on edge or mobile systems as a possible use.
Levelmodel
Classreasoning
Parameters3.85B
Context64,000
Released19 Jun 2025
Opennessopen weights permissive
Runs onmobile or NPU
Statusactive
From the model's own page; who made each picture and where it comes from are listed on the Image Credits page.
Benchmark Results
vendor-reported
AIME 2025 2025math
vendor-reported
33.59
050100
Score in %. Higher is better.
AIME24 2024math
vendor-reported
52.29
050100
Score in %. Higher is better.
GPQA Diamondknowledge and reasoning
vendor-reported
45.08
050100
Score in %. Higher is better.
MATH500math
vendor-reported
92.45
050100
Score in %. Higher is better.
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).
Identity
- Developer
- Microsoft AI
- Family
- phi-4
- Released
- 19 Jun 2025 sourcesregistry date
- Weights published
- –
- Record checked
- 10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.
Architecture and Size
- Architecture
- hybrid SambaY decoder-hybrid-decoder: state-space layers, Gated Memory Unit and Differential Attention
- Total parameters
- 3.85B sources
- Active parameters
- –
- Checkpoint parameters
- –
- Context length
- 64,000 tokens
- Input
- text
- Output
- text
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).
Openness and Access
- Openness
- open weights permissive
- Licence
- MIT License sources
- Access routes
- open weights download, API
Where this comes from
The licence file governs over registry tags and aggregator labels.
Deployment
- Runs on
- mobile or NPU
- Footprint
- Up to 10x decoding throughput over Phi-4-mini-reasoning (2K prompt, 32K generation), measured with vLLM on a single A100-80GB (not edge hardware). Quantized footprint and on-device latency unverified
- Runtimes
- not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.
Indonesia
- Indonesian support
- not stated
- Regional languages
- not in the notes
- Jakarta hosting
- unknown
- Languages
- not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.
Lineage
The notes record no lineage for this model.
Industry Use Cases
No industry use case links this model.
Links
DeveloperMicrosoft AIbig tech, USA
Microsoft AIbig tech, USAModelPhi-4-mini-flash-reasoningMicrosoft AI, reasoning
Phi-4-mini-flash-reasoningMicrosoft AI, reasoningUse CasesIndustrial Copilots and Maintenance AssistantsIndustrialChinese Medical LLMsHealthcareIndonesian and Regional-Language LLMsPublic SectorTelecom Domain LLMs and BenchmarksTelecomUse cases of models in the same R&D topic; none is recorded for this model itself.
Who Runs It
PT Pertamina Geothermal EnergyIn production, 14 Jul 2026
Siemens Erlangen electronics factoryPlanned, 6 Jan 2026
ADNOC OffshoreIn production, 9 Sep 2025
Aker BPIn production, 9 Sep 2025+10 moreDeployments of these use cases; none names this model.
PT Pertamina Geothermal EnergyIn production, 14 Jul 2026Siemens Erlangen electronics factoryPlanned, 6 Jan 2026ADNOC OffshoreIn production, 9 Sep 2025Aker BPIn production, 9 Sep 2025+10 moreDeployments of these use cases; none names this model.Models 1
Organisations 1
Sources
- huggingface.co/microsoft/Phi-4-mini-flash-reasoning https://huggingface.co/microsoft/Phi-4-mini-flash-reasoning
- huggingface.co/api/models/microsoft/Phi-4-mini-flash-reasoning https://huggingface.co/api/models/microsoft/Phi-4-mini-flash-reasoning
- microsoft / phi-4-mini-flash-reasoning docs.api.nvidia.com https://docs.api.nvidia.com/nim/reference/microsoft-phi-4-mini-flash-reasoning
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Phi-4-Mini-Flash-Reasoning", https://aurion.semai.id/models/microsoft-phi-4-mini-flash-reasoning/, records of October 2026.