Model Follower.

Up to date

Rebuilt
2026-10-25 06:54 UTC
Release radar
117 on watch
Hub snapshot
2026-10-25
Re-checked
daily · every number sourced and dated

Models › Phi-3 family

Phi-3-mini-128k-instruct

Hugging Face ↗ Follow changes

microsoft MIT out 3 years ago

3.82B parameters, text in, text out, 128K tokens of context.

Verified 2026-10-25 Rank #210 in LLMs · score 88.4 method v1 Sources 3 Downloads 30d 246.5k · #236 of 619 tracked

Can I run it?2.3GB at INT4Estimatefits a 16 GB cardmethodCommunity: Phi 3 Mini 128K Instruct by microsoft — VRAM 7.7GB, 128K context | LLM Explorer →
Can I use it?MITPermissivelicence file2026-10-24
Do people use it?246.5kdownloads, 30 days#236 of 619 we track · top 38%+3 likes, downloads +20% this week1,711 likesHugging Face2026-10-24

Worth reading

Picked by rule from 23 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Most discussed · 411 HN pts

Phi-3 Technical Report

Phi-3 models achieve high performance rivaling Mixtral 8x7B by training on a scaled-up dataset of heavily filtered publicly available web data and synthetic data.

arxiv.org 2024-04-23 411 HN pts · 130 comments 463 words about the Phi 3 family family
Latest · 22 months ago

Exploring Microsoft's Phi-3-Mini and its integration with tool like Ollama

Microsoft's Phi-3-Mini is a 3.8 billion-parameter language model designed for efficiency on devices with limited resources like notebooks and smartphones.

  • Organizations needing full-scale enterprise solutions should consider deploying phi-3.5-mini on Azure for production-grade stability.
  • Pieces CLI can access your pastebin to store code snippets generated by the local phi-3-mini model.
pieces.app 2024-12-26 84 HN pts · 31 comments 1.2k words about the Phi 3 family family

23 best of 23 tracked

The 2 picks above, plus these 21. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

  1. CodeOllama v0.1.33 with Llama 3, Phi 3, and Qwen 110Bgithub.com · 2024-04-28 · 192 HN pts
  2. CodeMicrosoft Phi-3 Cookbookgithub.com · 2024-05-21 · 152 HN pts
  3. ArticlePhi-3: Redefining what's possible with SLMsazure.microsoft.com · 2024-04-23 · 16 HN pts · official
  4. PaperPhi-3 Technical Report: A Highly Capable Language ...arxiv.org · undated
  5. ArticlePhi-3-Mediumazure.microsoft.com · 2024-05-21 · 6 HN pts · official
  6. ArticlePhi-3 Technical a Highly Capable Language Model Locally on Your Phonehuggingface.co · 2024-04-23 · 7 HN pts
  7. DiscussionMicrosoft Updated Phi-3twitter.com · 2024-07-02 · 35 HN pts
  8. ArticleTiny but mighty: The Phi-3 small language models with big potentialnews.microsoft.com · 2024-04-23 · 6 HN pts · official
  9. ArticleBuilding Intelligent Applications with Local RAG in .NET and Phi-3techcommunity.microsoft.com · 2024-07-02 · official
  10. ArticleMicrosoft launches Phi-3, its smallest AI model yettheverge.com · 2024-04-23
  11. ArticleThe latest major open LLM releases: Mixtral, Llama 3, Phi-3, and OpenELMmagazine.sebastianraschka.com · 2024-05-12
  12. ArticleLocal iOS/macOS phi-3-mini-4k-instructapps.apple.com · 2024-05-14
  13. ArticleMicrosoft releases powerful new Phi-3.5 models, beating Google, OpenAI and moreventurebeat.com · 2024-08-21
  14. PaperPhi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonearxiv.org · 2024-08-30
  15. CodeShow HN: Phi-3-MLX – Language and Vision Models for Apple Silicongithub.com · 2024-07-19
  16. ArticlePhi-3-vision in 50 lines of C# with ONNX Runtime GenAInietras.com · 2024-06-05
  17. ArticleAfter Llama3 here we have Phi-3: Small Language Models beating LLMsdidyouknowbg8.wordpress.com · 2024-04-23
  18. ArticlePhi-3-mini-128k-instruct | Model Catalog | Microsoft Foundryai.azure.us · undated
  19. ArticlePhi 3 Mini 128K Instruct by microsoft — VRAM 7.7GB, 128K context | LLM Explorerllm-explorer.com · undated
  20. ArticleTrain Phi-3 Visionwandb.ai · 2024-06-01
  21. ArticleTiny LLM Reviews: Phi 3 3.8bkyle.works · 2024-04-30

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

No API provider serves it, so its downloads are people running it themselves: 87 quantized builds and 100 Spaces.

  • As a service0no API provider listed
  • On own hardware87quantized builds
  • Built on44944 fine-tunes · 392 adapters · 13 merges
  • In apps100+Hugging Face Spaces

Top Spaces running it: jbilcke-hf/tikslop ♥ 305 · allenai/WildBench ♥ 232 · allenai/ZebraLogic ♥ 95 · microsoft/MInference ♥ 61

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Fits a 16 GB card at INT4

About 2.3 GB at INT4, 4.6 GB at FP8, 9.2 GB at BF16: weights only, so leave room for the KV cache.

16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 2.3 GB of weights (estimate; KV cache and activations extra)INT42.3 GBFP8: about 4.6 GB of weights (estimate; KV cache and activations extra)FP84.6 GBBF16: about 9.2 GB of weights (estimate; KV cache and activations extra)BF169.2 GB
16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 2.3 GB of weights (estimate; KV cache and activations extra)INT42.3 GBFP8: about 4.6 GB of weights (estimate; KV cache and activations extra)FP84.6 GBBF16: about 9.2 GB of weights (estimate; KV cache and activations extra)BF169.2 GB

Estimate params × bytes per param × 1.2 · how we estimate

The story so far

Weights, every dated article, and repo commits on one rail.

Most coverage came in the first two weeks

Teal is the weights (2024-04-22), amber triangles are 9 dated articles (each links to it), grey ticks are repo commits.

MayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2025-12-10: 2 repo commits2025-03-02: 1 repo commit2024-08-20: 1 repo commit2024-07-01: 2 repo commits2024-06-14: 1 repo commit2024-05-31: 2 repo commits2024-05-30: 1 repo commit2024-05-22: 1 repo commit2024-05-21: 1 repo commit2024-05-02: 2 repo commits2024-04-29: 2 repo commits2024-04-26: 3 repo commits2024-04-25: 2 repo commits2024-04-22: 2 repo commitsPhi-3 Technical Report · 2024-04-23Ollama v0.1.33 with Llama 3, Phi 3, and Qwen 110B · 2024-04-28Microsoft Phi-3 Cookbook · 2024-05-21Phi-3: Redefining what's possible with SLMs · 2024-04-23Exploring Microsoft's Phi-3-Mini and its integration with tool like Ollama · 2024-12-26Phi-3-Medium · 2024-05-21Phi-3 Technical a Highly Capable Language Model Locally on Your Phone · 2024-04-23Microsoft Updated Phi-3 · 2024-07-02Tiny but mighty: The Phi-3 small language models with big potential · 2024-04-23Weights 2024-04-22 · 2024-04-22Weights 2024-04-22now
MayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2025-12-10: 2 repo commits2025-03-02: 1 repo commit2024-08-20: 1 repo commit2024-07-01: 2 repo commits2024-06-14: 1 repo commit2024-05-31: 2 repo commits2024-05-30: 1 repo commit2024-05-22: 1 repo commit2024-05-21: 1 repo commit2024-05-02: 2 repo commits2024-04-29: 2 repo commits2024-04-26: 3 repo commits2024-04-25: 2 repo commits2024-04-22: 2 repo commitsPhi-3 Technical Report · 2024-04-23Ollama v0.1.33 with Llama 3, Phi 3, and Qwen 110B · 2024-04-28Microsoft Phi-3 Cookbook · 2024-05-21Phi-3: Redefining what's possible with SLMs · 2024-04-23Exploring Microsoft's Phi-3-Mini and its integration with tool like Ollama · 2024-12-26Phi-3-Medium · 2024-05-21Phi-3 Technical a Highly Capable Language Model Locally on Your Phone · 2024-04-23Microsoft Updated Phi-3 · 2024-07-02Tiny but mighty: The Phi-3 small language models with big potential · 2024-04-23Weights 2024-04-22 · 2024-04-22Weights 2024-04-22now

What people built on it

536 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

536 derivatives on the Hub, most of them adapters

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

87 quantized (Hugging Face model tree)44 fine-tunes (Hugging Face model tree)13 merges (Hugging Face model tree)392 adapters (Hugging Face model tree)87 quantized44 fine-tunes13 merges392 adapters
87 quantized (Hugging Face model tree)44 fine-tunes (Hugging Face model tree)13 merges (Hugging Face model tree)392 adapters (Hugging Face model tree)87 quantized44 fine-tunes13 merges392 adapters
GGUF bartowski/Phi-3.1-mini-128k-instruct-GGUF for llama.cpp, Ollama, LM Studio 163.8k downloads AWQ solidrust/Phi-3-mini-128k-instruct-AWQ for vLLM, SGLang 10.6k downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Benchmarks

20 rows

Benchmarks 20 rows

measured third-party leaderboard rowauthor-reported number from the model card or eval files (20)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.
BenchmarkValueSetting / notesReported byDate
author-reported Average37.3June 2024 Updatemodel author2026-10-07
author-reported GPQA Diamond · Reason through hard problems29.7June 2024 Updatemodel author2026-10-07
author-reported Instruction Extra Hard5.9June 2024 Updatemodel author2026-10-07
author-reported Instruction Hard5.2June 2024 Updatemodel author2026-10-07
author-reported JSON Structure Output60.1June 2024 Updatemodel author2026-10-07
author-reported MMLU · Reason through hard problems69.7June 2024 Updatemodel author2026-10-07
author-reported RULER · Read long documents92.4June 2024 Update 4Kmodel author2026-10-07
author-reported RULER · Read long documents91.1June 2024 Update 8Kmodel author2026-10-07
author-reported RULER · Read long documents90.8June 2024 Update 16Kmodel author2026-10-07
author-reported RULER · Read long documents87.9June 2024 Update 32Kmodel author2026-10-07
author-reported RULER · Read long documents79.8June 2024 Update 64Kmodel author2026-10-07
author-reported RULER · Read long documents65.6June 2024 Update 128Kmodel author2026-10-07
author-reported RULER · Read long documents84.6June 2024 Update Averagemodel author2026-10-07
author-reported RepoQA · Write code with agents85June 2024 Update Pythonmodel author2026-10-07
author-reported RepoQA · Write code with agents63June 2024 Update C++model author2026-10-07
author-reported RepoQA · Write code with agents72June 2024 Update Rustmodel author2026-10-07
author-reported RepoQA · Write code with agents93June 2024 Update Javamodel author2026-10-07
author-reported RepoQA · Write code with agents72June 2024 Update TypeScriptmodel author2026-10-07
author-reported RepoQA · Write code with agents77June 2024 Update Averagemodel author2026-10-07
author-reported XML Structure Output52.9June 2024 Updatemodel author2026-10-07

What the card says

2 cited sentences

Summary generated, cited

Phi-3-mini-128k-instruct is a 3.82B parameter text-generation model from the Phi-3 family.[1][2] The model supports a context length of 128K tokens and is released under the MIT license.[3][4]

  1. parameters human
  2. “Phi-3 family”
  3. “context length (in tokens) that it can support”
  4. license

Generated 2026-10-07 15:05 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.

Use-case evidence

5 use cases

What it is good at 5 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Facts

21 fields

Listed in LLMs

Facts

Pipeline tagtext-generation
Librarytransformers
ArchitecturesPhi3ForCausalLM
Model typephi3
Dtypebfloat16
Params by dtypeBF16: 3,821,079,552
Size bandM
MoEno
Context length131,072 (config.json max_position_embeddings)
Languagesen
LicenseMIT (mit)
Hub snapshot2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · Phi3ForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Phi3ForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for Phi-3-mini-128k-instruct: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)9.2 GB · Single GPU (<24 GB)
FP8 (1 byte/param)4.6 GB · Single GPU (<24 GB)
INT4 (0.5 bytes/param)2.3 GB · Edge (<4 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.

All variants

40 variants

Lineage

Original model Family: microsoft/Phi-3 536 derivatives on the Hub

merge: 13quantized: 87finetune: 44adapter: 392

Top variants (40 shown, ranked by downloads)

Relation
Quant
40 rows
Variant Relation Quant Params Band License Author Likes Downloads
OpenGVLab/InternVL2-4B merge — 4.15B M mit OpenGVLab 57 597.4k
bartowski/Phi-3.1-mini-128k-instruct-GGUF quantized GGUF — — mit bartowski 36 163.8k
lmstudio-community/Phi-3.1-mini-128k-instruct-GGUF quantized GGUF — — mit lmstudio-community 7 87.4k
OpenGVLab/Mini-InternVL-Chat-4B-V1-5 merge — 4.15B M mit OpenGVLab 62 50.5k
legraphista/Phi-3-mini-128k-instruct-IMat-GGUF quantized GGUF — — mit legraphista 0 29.1k
second-state/Phi-3-mini-128k-instruct-GGUF quantized GGUF — — mit second-state 2 24k
qwp4w3hyb/Phi-3-mini-128k-instruct-iMat-GGUF quantized GGUF — — mit qwp4w3hyb 0 23k
gaianet/Phi-3-mini-128k-instruct-GGUF quantized GGUF — — mit gaianet 1 12.9k
solidrust/Phi-3-mini-128k-instruct-AWQ quantized AWQ 3.82B M — solidrust 1 10.6k
professorf/phi-3-mini-128k-f16-gguf quantized gguf — — mit professorf 5 10.1k
mradermacher/Phi-3-mini-128k-instruct-i1-GGUF quantized GGUF — — mit mradermacher 0 5.1k
tensorblock/Phi-3-mini-128k-instruct-GGUF quantized GGUF — — mit tensorblock 0 5k
calcuis/phi3 quantized GGUF — — mit calcuis 0 5k
OpenVINO/Phi-3-mini-128k-instruct-int4-ov quantized int4 — — mit OpenVINO 3 4.6k
OEvortex/EMO-phi-128k finetune — 3.82B M mit OEvortex 10 4k
AlessandroW/Phi-3-mini-128k-instruct-gguf quantized gguf — — mit AlessandroW 1 3.2k
RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a8 quantized -quant 3.82B M mit RedHatAI 0 2.8k
mradermacher/Phi-3-mini-128k-instruct-GGUF quantized GGUF — — mit mradermacher 1 2.7k
Ffftdtd5dtft/Phi-3-mini-128k-instruct-IQ2_XXS-GGUF quantized GGUF — — mit Ffftdtd5dtft 1 2.6k
BenevolenceMessiah/Phi-3-mini-128k-instruct-GGUF quantized GGUF — — mit BenevolenceMessiah 0 2.6k
OpenVINO/Phi-3-mini-128k-instruct-int8-ov quantized int8 — — mit OpenVINO 5 2.2k
justinthelaw/Phi-3-mini-128k-instruct-4bit-128g-GPTQ quantized -4bit 3.82B M apache-2.0 justinthelaw 1 1.9k
Ffftdtd5dtft/Phi-3-mini-128k-instruct-Q2_K-GGUF quantized Q2_K — — mit Ffftdtd5dtft 0 1.9k
CHE-72-ZLab/Microsoft-Phi3-14B-Instruct128K-GGUF quantized GGUF — — mit CHE-72-ZLab 0 1.8k
leliuga/Phi-3-mini-128k-instruct-bnb-4bit quantized -bnb 3.93B M other leliuga 1 1.5k
OpenVINO/Phi-3-mini-128k-instruct-fp16-ov finetune -fp16 — — mit OpenVINO 0 1.2k
newsletter/Phi-3-mini-128k-instruct-Q6_K-GGUF quantized Q6_K — — mit newsletter 0 1.1k
paulilioaica/MixtureOfPhi3 finetune — 6.24B M apache-2.0 paulilioaica 9 630
mlc-ai/Phi-3-mini-128k-instruct-q4f16_1-MLC quantized — — — — mlc-ai 1 951
DhruvalLabs/Phi-3-mini-128k-instruct-GGUF quantized GGUF — — mit DhruvalLabs 0 920
gate369/Phi-3-mini-128k-instruct-IQ4_XS-GGUF quantized GGUF — — mit gate369 0 893
RedHatAI/Phi-3-mini-128k-instruct-quantized.w8a16 quantized -quant 3.82B M mit RedHatAI 0 851
CitrusBoy/FinetunedModelV2.0 adapter — 3.88B M mit CitrusBoy 0 839
CHE-72/Phi-3-mini-128k-instruct-Q6_K-GGUF quantized Q6_K — — mit CHE-72 0 823
Muhammad2003/Orpo-Phi3-3B-128K finetune — 3.82B M apache-2.0 Muhammad2003 3 617
praysimanjuntak/llava-phi3-3.8b-lora finetune — 4.14B M — praysimanjuntak 5 517
awels/maximusLLM-3b-128k-gguf quantized gguf — — mit awels 1 663
Ejafa/phi-3-mini-128k-instruct-simpo-lr-5e-07-gamma-1.5 finetune — 3.82B M mit Ejafa 0 685
Antonio88/TaliML-PHI3-128K-ITA-V.1.0.FINAL finetune — 3.82B M mit Antonio88 0 642
suraiy/microsoft-Phi-3-mini-128k-instruct-HQQ-4bit-smashed-Q4_K_M-GGUF quantized -4bit — — — suraiy 0 623

Repository changelog

23 commits

Repository changelog 23 commits

DateCommitSHA
2025-12-10 Update README.md (#104) f3c06ae
2025-12-10 Upload data_summary_card.md (#103) 17fb6f6
2025-03-02 Update readme 072cb75
2024-08-20 Update README.md a90b62a
2024-07-01 Minor updates for README d548c23
2024-07-01 Update model (#84) 10d25df
2024-06-14 Add the classifiers to the auto_map (#76) bb5bf1e
2024-05-31 Add attention_bias to make TGI work (#68) 5be6479
2024-05-31 Add link to Phi-3 vision ONNX models 6417fb9
2024-05-30 Update sample_finetune.py (#70) 3c014d5
2024-05-22 Update sample_finetune.py (#69) bbd531d
2024-05-21 Update README 30e8c65
2024-05-02 fix(tokenizer): Also updates tokenizer.json to prevent any mismatch. 8a362e7
2024-05-02 fix(tokenizer_config): Adjusts `rstrip` of special tokens. (#54) 82eebd7
2024-04-29 Update README.md 39d6675
2024-04-29 fix(root): Replaces system by user to improve generation experience. fc313b0
2024-04-26 Update README.md 810c2e6
2024-04-26 Update sample_finetune.py (#42) 0693e0b
2024-04-26 chore(readme): Adds a proper chat example to widget. 49f911d
2024-04-25 fix(readme): Adds information about placeholder tokens. f10fb29
2024-04-25 fix(tokenizer_config): Ensures model_max_length corresponds to actual maximum number. 48788d7
2024-04-22 chore(root): Initial files upload. 7de6f02
2024-04-22 initial commit f310673

Articles and coverage

23 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.

Official3

Phi-3: Redefining what's possible with SLMs
azure.microsoft.com2024-04-23officialauthorityabout the Phi 3 familyHN 16 pts · 1 comments1,074 words
Microsoft introduces Phi-3, a family of open small language models including the upcoming Phi-3-small (7B) and Phi-3-medium (14B). Phi-3-mini is a 3.8B parameter model available on Azure AI Studio, Hugging Face, and Ollama with a context window of up to 128K tokens. The models have lower capacity to retain facts on factual knowledge benchmarks due to their smaller size.written, reviewed, cited
  • Phi-3-mini can be used on-device, especially when further optimized with ONNX Runtime for cross-platform availability.
Phi-3-Medium
azure.microsoft.com2024-05-21officialauthorityabout the Phi 3 familyHN 6 pts · 1 comments1,243 words
Tiny but mighty: The Phi-3 small language models with big potential
news.microsoft.com2024-04-23officialauthorityabout the Phi 3 familyHN 6 pts · 0 comments

Analysis6

Phi-3 Technical Report
arxiv.org2024-04-23authorityHN front pageabout the Phi 3 familyHN 411 pts · 130 comments463 words
Phi-3 models achieve high performance rivaling Mixtral 8x7B by training on a scaled-up dataset of heavily filtered publicly available web data and synthetic data. This approach enables deployment locally on your phone despite small parameter counts like the 3.8 billion in phi-3-mini. The phi-3.5-MoE variant reaches 78% on MMLU with 6.6 billion active parameters, while phi-3.5-Vision handles multi-image prompts.written, reviewed, cited
Ollama v0.1.33 with Llama 3, Phi 3, and Qwen 110B
github.com2024-04-28authorityHN front pageabout the Phi 3 familyHN 192 pts · 64 comments38 words
Microsoft Phi-3 Cookbook
github.com2024-05-21authorityHN front pageabout the Phi 3 familyHN 152 pts · 57 comments978 words
Microsoft released the Phi family of open-source small language models, which are the most capable and cost-effective SLMs available. The repository includes 50+ language translations that significantly increase download size, requiring sparse checkout to clone without them. Specific samples include Phi-4-Reasoning-Vision-15B for reasoning-vision tasks, and the models support deployment on cloud or edge devices. Developers should integrate Azure AI Content Safety to detect harmful user-generated and AI-generated content in applications.written, reviewed, cited
  • Developers can use sparse checkout commands to clone the repository without downloading the translation files.
  • The project contains specific samples for Phi-4-Mini-Flash-Reasoning-Instruct for math demonstrations.
Exploring Microsoft's Phi-3-Mini and its integration with tool like Ollama
pieces.app2024-12-26about the Phi 3 familyHN 84 pts · 31 comments1,213 words
Microsoft's Phi-3-Mini is a 3.8 billion-parameter language model designed for efficiency on devices with limited resources like notebooks and smartphones. It excels in reasoning and coding tasks, offering performance comparable to larger models such as GPT-3.5 while requiring less computational power. Variants like phi-3-mini-128k-instruct support extensive context windows, though the model currently lacks context window overflow handling and may reply with gibberish until a future ONNX release fixes the issue.written, reviewed, cited
  • Organizations needing full-scale enterprise solutions should consider deploying phi-3.5-mini on Azure for production-grade stability.
  • Pieces CLI can access your pastebin to store code snippets generated by the local phi-3-mini model.
Phi-3 Technical Report: A Highly Capable Language ...
arxiv.orgauthorityabout the Phi 3 family6,096 words
Phi-3 Technical a Highly Capable Language Model Locally on Your Phone
huggingface.co2024-04-23authorityabout the Phi 3 familyHN 7 pts · 7 comments1,073 words

Community1

Microsoft Updated Phi-3
twitter.com2024-07-02about the Phi 3 familyHN 35 pts · 15 comments
All coverage tracked for this model (23) — append-only, nothing is ever removed
TitleSourcePublishedFirst seenSignals
Phi-3 Technical Report: A Highly Capable Language ...arxiv.org—2026-10-07exa
Phi-3-mini-128k-instruct | Model Catalog | Microsoft Foundryai.azure.us—2026-10-07exa
Phi 3 Mini 128K Instruct by microsoft — VRAM 7.7GB, 128K context | LLM Explorerllm-explorer.com—2026-10-07exa
Exploring Microsoft's Phi-3-Mini and its integration with tool like Ollamapieces.app2024-12-262026-10-07hn
Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phonearxiv.org2024-08-302026-10-07exa
Microsoft releases powerful new Phi-3.5 models, beating Google, OpenAI and moreventurebeat.com2024-08-212026-10-07hn
Show HN: Phi-3-MLX – Language and Vision Models for Apple Silicongithub.com2024-07-192026-10-07hn
Building Intelligent Applications with Local RAG in .NET and Phi-3techcommunity.microsoft.com official2024-07-022026-10-07hn
Microsoft Updated Phi-3twitter.com2024-07-022026-10-07hn
Phi-3-vision in 50 lines of C# with ONNX Runtime GenAInietras.com2024-06-052026-10-07hn
Train Phi-3 Visionwandb.ai2024-06-012026-10-07hn
Microsoft Phi-3 Cookbookgithub.com2024-05-212026-10-07hn
Phi-3-Mediumazure.microsoft.com official2024-05-212026-10-07hn
Local iOS/macOS phi-3-mini-4k-instructapps.apple.com2024-05-142026-10-07hn
The latest major open LLM releases: Mixtral, Llama 3, Phi-3, and OpenELMmagazine.sebastianraschka.com2024-05-122026-10-07hn
Tiny LLM Reviews: Phi 3 3.8bkyle.works2024-04-302026-10-07exa
Ollama v0.1.33 with Llama 3, Phi 3, and Qwen 110Bgithub.com2024-04-282026-10-07hn
After Llama3 here we have Phi-3: Small Language Models beating LLMsdidyouknowbg8.wordpress.com2024-04-232026-10-07hn
Phi-3: Redefining what's possible with SLMsazure.microsoft.com official2024-04-232026-10-07hn, exa
Tiny but mighty: The Phi-3 small language models with big potentialnews.microsoft.com official2024-04-232026-10-07hn
Phi-3 Technical a Highly Capable Language Model Locally on Your Phonehuggingface.co2024-04-232026-10-07hn
Microsoft launches Phi-3, its smallest AI model yettheverge.com2024-04-232026-10-07hn
Phi-3 Technical Reportarxiv.org2024-04-232026-10-07hn

Similar models

3 models

Similar models

  • Qwen3-4B · 4.02B
    same size band (M), same task (text-generation), more providers, permissive license
  • Qwen3-4B-Instruct-2507 · 4.02B
    same size band (M), same task (text-generation), more providers, permissive license
  • Phi-3.5-mini-instruct · 3.82B
    same size band (M), same task (text-generation), permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:37 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:05 UTC · Last verified 2026-10-25

Follow Phi-3-mini-128k-instruct

Watch this model

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.