Headless Thrones

Microsoft · Official-source model record

Phi-4-reasoning

Phi

Listed by publisherChecked Oct 23, 2026
Editorial (0) ↓Download model brief (.md)

Publisher-sourced inventory. Specifications come from official documentation and model cards. “Not documented” means we have no verified value. This inventory is separate from our editorial reviews.

Coverage & sources ↗

Current catalog answers

What to know about Phi-4-reasoning

AI-compiled from linked model documentation. This is a catalog reference, not an independent performance review. Catalog checked October 23, 2026.

Which Phi-4-reasoning does this page describe?

Phi-4-reasoning is a model record from Microsoft in the Phi family. Source

A separate version identifier is not documented in this record; the model name alone does not establish which snapshot a service runs.

How can you access Phi-4-reasoning?

The catalog documents Open weights access for Phi-4-reasoning. Source

As of the catalog check, this version was listed without confirmed general availability. Source

Open weights means the model weights are distributed for use under their stated terms. It does not establish a hosted API, an open-source license, or a hardware requirement.

What are Phi-4-reasoning's context and output limits?

The documented context window is 32,000 tokens. Source

A separate maximum-output value is not established in this record. The context or input figure should not be used as an output allowance.

Context, input and output limits describe different constraints. Do not add the figures together to infer a larger window. Check the endpoint's accounting for conversation history, system instructions, tools, media and generated tokens; a documented maximum is not a guarantee of answer quality or a measured usable capacity.

What dates are documented for Phi-4-reasoning?

Release date: April 30, 2025. Source This is a release milestone, not a knowledge or training cutoff.

Neither a knowledge cutoff nor a training-data cutoff is established here. A release date, repository upload or filename cannot supply those missing dates.

What can Phi-4-reasoning receive and generate?

Documented inputs: text. Source

Documented outputs: text. Source

Input support and output generation are separate capabilities. Reading an image does not establish image generation; a text output listing does not establish speech output. File sizes, formats and media limits depend on the documented implementation.

What license is recorded for Phi-4-reasoning?

The recorded license label is mit. Source

Read the linked license and associated use terms for this exact version before redistribution, commercial deployment or fine-tuning. A label alone does not establish every permission, restriction or obligation, and it does not determine the terms of a separate hosted service.

What is documented about the Phi-4-reasoning checkpoint?

This evidence refers to microsoft/Phi-4-reasoning at recorded revision 1de18ec97600877ce63dbf60c73b998da99f0195, checked October 23, 2026. Repository metadata

Safetensors metadata records 14,659,507,200 stored parameters. This is a stored-weight count, not necessarily the number activated per token, a download size or a GPU-memory requirement. Repository metadata

Recorded architecture class: Phi3ForCausalLM. Repository metadata

Repository task tag: text-generation. A task tag identifies a distribution category; it does not measure performance on that task. Repository metadata

Listed library: transformers. Check that library's version and model support before loading the weights. Repository metadata

Repository language labels: en. These labels are not independent evaluations of language quality. Repository metadata

Repository access gate: false. This records the repository's gate setting when checked; licensing and use terms still apply. Repository metadata

What the configuration file says

The pinned config.json contains these values. Pinned configuration

max_position_embeddings
32,768
torch_dtype
bfloat16
num_attention_heads
40
num_key_value_heads
10
num_hidden_layers
40
hidden_size
5,120

Configuration values describe the checkpoint and its implementation. In particular, max_position_embeddings, sliding_window and rope_scaling do not independently establish an API context limit, tested usable length or hardware requirement. Runtime, quantization, cache, batch size and input length affect deployment behavior.

Read the retrieved model card for setup and qualifications. Model card

Keep the exact version in view

Other Phi models from Microsoft.

These are separate records in the same publisher family. A shared family name does not establish equal capabilities or interchangeable licenses.

The documented record

Specifications, with their sources.

A missing value remains unknown. Input limits, output limits and the total context window are separate measurements.

Publisher
MicrosoftSource ↗
Family
PhiSource ↗
Version
Not documented
Availability
Listed by publisherSource ↗
Access
Open weightsSource ↗
Inputs
TextSource ↗
Outputs
TextSource ↗
Context window
32,000 tokensSource ↗
Maximum input
Not documented
Maximum output
Not documented
Knowledge cutoff
Not documented
Training-data cutoff
Not documented
Training completed
Not documented
Release date
Apr 30, 2025Release date · day precisionSource ↗
License
mitSource ↗
Input price / 1M tokens
Not documented
Output price / 1M tokens
Not documented

Read this with the numbers.

Official publisher model repository; listed status confirms a catalog entry, not a verified hosted API service. Context is the publisher’s nominal 32k token figure (K/M normalized decimally), not a tested request limit. Publisher repository evidence (2026-10-23): Publisher task label: text-generation. Published architecture metadata: Phi3ForCausalLM. Repository library label: transformers. Model-card language labels: en; these labels are not evaluated proficiency scores. Published safetensors metadata counts 14,659,507,200 stored parameters; this is not an active-parameter count for mixture-of-experts models. Published configuration: max_position_embeddings=32768; dtype=bfloat16; 40 hidden layers. Checkpoint configuration is not a verified hosted serving limit; context and output limits remain separate documented fields.

Model identifiers and aliases 1
  • microsoft/Phi-4-reasoning

Follow the evidence

Official sources.

  1. Phi-4-reasoning — official model card ↗huggingface.co
  2. Phi-4-reasoning: official publisher repository metadata ↗huggingface.co
  3. Phi-4-reasoning: publisher model card at revision 1de18ec97600 ↗huggingface.co
  4. Phi-4-reasoning: published checkpoint configuration at revision 1de18ec97600 ↗huggingface.co

From the Headless Thrones newsroom

Editorial about Phi-4-reasoning

No published articles are linked to this exact model yet. New coverage will appear here as it is published.

Your starting point

Phi-4-reasoning

Microsoft · Open weights

Browse all models ↗
Inputs
TextSource ↗
Outputs
TextSource ↗
Context window
32,000 tokensSource ↗
Knowledge cutoff
Not documented

A place to begin comparing

1,133 models to consider.

Matches must share an access route, support every documented input type of your starting model, and return text. When both publish a context limit in the same field, the candidate must fall between half and twice that limit. Unknown or differently defined limits remain as broader matches, after verified near-context matches. Within each group, recent documented releases appear first. These criteria do not establish equal quality. The full current inventory is considered.

Microsoft

Phi-4-reasoning-vision-15B

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Mistral AI

Voxtral-Small-24B-2507

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Microsoft

Phi-4-mini-flash-reasoning

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Microsoft

Phi-4-reasoning-plus

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Microsoft

phi-4

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Arcee AI

Arcee-Blitz

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Arcee AI

Arcee-VyLinh

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Arcee AI

Virtuoso-Lite

Text input; text output · Shared access: Open weights · Documented context within ½–2×

ByteDance Seed

Seed-Coder-8B-Base

Text input; text output · Shared access: Open weights · Documented context within ½–2×

ByteDance Seed

Seed-Coder-8B-Instruct

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Cohere

aya-vision-32b

Text input; text output · Shared access: Open weights · Documented context within ½–2×

Cohere

North-Small-Translate-1.0

Text input; text output · Shared access: Open weights · Documented context within ½–2×

A record you can follow

The source belongs beside the fact.

Official sources

Model records cite the developer’s documentation, release notes and model cards. A listing alone is not a tested service or a recommendation.

Dates mean different things

Knowledge cutoff, training-data cutoff, training completion and release date are separate fields. A model’s name or repository upload date does not establish any of them.

Limits need context

Some publishers use rounded K/M token figures. Deployment, reasoning settings and regional access can change usable limits. Read the source before choosing.

Coverage and last check

Independent editorial model library assembled from primary vendor documentation and official publisher model cards. This is a dated, auditable inventory of the sources reviewed, not a claim to every model worldwide. First-party language-model inventory for OpenAI, Anthropic, Google Gemini/Gemma, xAI, Meta and Reflection. Includes vendor-listed legacy/retired identities with explicit statuses; official pretrained and instruction-tuned checkpoints remain distinct. Excludes embeddings, moderation/guard models, image/audio/video-only generation, community fine-tunes, speculative draft helpers, format/runtime/quantization duplicates, and PaliGemma dataset-specific benchmark fine-tunes. Availability means documented vendor API or downloadable official weights, not guaranteed account eligibility. Unknown specifications and dates remain null. Rounded K/M limits preserve vendor granularity; exact integers are used only where documented. No OpenRouter facts are used. Primary-source inventory of language-capable models from Alibaba/Qwen, DeepSeek, Moonshot, MiniMax, Z.ai, ByteDance Seed, Tencent, Baidu, Xiaomi, StepFun, 01.AI, Shanghai AI Laboratory, Baichuan and OpenBMB. Includes full reachable official public weight inventories plus documented hosted API models and explicit retirements. Embeddings, reward/classifier/router/tokenizer/draft-head components, standalone OCR/document converters and image/video/audio-only generators are excluded; multimodal language models retained. Official packaging variants are folded into aliases where a canonical listing was identifiable; remaining research/base/SFT variants retain exact identities. Not a guarantee every historical/private/regional model is available. Unknown specifications remain null; K/M API display quantities use decimal normalization where explicitly noted, rather than implying verified exact serving ceilings. Model-name dates and repository creation timestamps are never used as release/training dates. Open weights does not mean unrestricted or OSI-open-source licensing. Some model-card bodies were unreachable; their public inventory metadata is retained and disclosed. Official Mistral, Cohere, AI21, NVIDIA Nemotron, Microsoft Phi/MAI, Amazon Nova, IBM Granite, AllenAI OLMo/Molmo, TII Falcon, Nous Research, Arcee and Essential AI model inventory checked 2026-10-23. Includes base, instruction, reasoning, coding and text-producing multimodal models, including legacy API versions. Excludes embeddings/rerankers/reward and safety classifiers, image/audio/video-only generation, robotics-only models, duplicate quantization/runtime packages, adapters, intermediate checkpoint experiments, and third-party mirrors. Arcee coverage focuses on its named product families rather than exploratory merge repositories. Repository listing means listed, not independently verified API availability. K/M figures are nominal decimal normalizations when the source supplies shorthand; deployment-specific exact token limits may differ. Unverified values remain null. This is a source-backed inventory, not a complete claim-level independent editorial review. Supplemental official Liquid AI, Reka and Inception generative language catalog. Excludes media-only generators and decision/classification-only models. No assumed training dates.

Catalog edition updated . Source links are provided on each model record.

View the full coverage report and discovery sources ↗