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 › granite family

granite-docling-258M

Hugging Face ↗ Follow changes

ibm-granite Apache 2.0 out 17 months ago

258M parameters, images and text in, text out, 8K tokens of context.

Verified 2026-10-25 Rank #261 in LLMs · score 85.5 method v1 Sources 3 Downloads 30d 295.3k · #220 of 619 tracked

Changed this week: 1 new quantized build on the Hub (16 in all).2026-10-24

Can I run it?0.2GB at INT4Estimatefits a 16 GB cardmethod
Can I use it?Apache 2.0Permissive; patent grantlicence file2026-10-24
Do people use it?295.3kdownloads, 30 days#220 of 619 we track · top 36%+4 likes, downloads +1% this week1,284 likesHugging Face2026-10-24

Worth reading

Picked by rule from 19 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Most discussed · 328 HN pts

Granite 4.1: IBM's 8B Model Matching 32B MoE

IBM released Granite 4.1, a family of open source models trained on 15 trillion tokens with a rigorous data quality pipeline.

  • The 3B model extends context only to 128K tokens, not the 512K available to larger variants.
  • Phase 5 extended the context window eventually to 512K tokens for the 8B and 30B models.
firethering.com 2026-04-30 328 HN pts · 209 comments 1.9k words about the granite family family

19 best of 19 tracked

The 2 picks above, plus these 17. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

  1. ArticleThe IBM Granite 4.1 family of modelsresearch.ibm.com · 2026-04-30 · 88 HN pts · official
  2. ArticleIBM Granite-Docling: End-to-end document understandingibm.com · undated · official
  3. CodeIBM Granite: A Family of Open Foundation Models for Code Intelligencegithub.com · 2024-05-07 · 252 HN pts
  4. ArticleIBM Granite 3.0: open enterprise modelsibm.com · 2024-10-21 · 39 HN pts · official
  5. ArticleOpen sourcing IBM's Granite code modelsresearch.ibm.com · 2024-05-07 · 18 HN pts · official
  6. ArticleIBM Granite 4.0ibm.com · 2025-10-02 · 10 HN pts · official
  7. Article'Western Qwen': IBM Wows with Granite 4 LLM Launch and Hybrid Mamba/Transformerventurebeat.com · 2025-10-03 · 83 HN pts
  8. ArticleIBM Research – Granite Foundation Models Technical Reportibm.com · 2023-11-29 · 4 HN pts · official
  9. ArticleIBM open-sources its Granite AI models – and they mean businesszdnet.com · 2024-05-13
  10. ArticleFun Local LLM Comparisons with Gemma, Granite, and Qwenekorbia.com · 2026-05-28
  11. CodeShow HN: A local-first financial auditor using IBM Granite, MCP, and SQLitegithub.com · 2026-01-01
  12. ArticleGranite 4.0 Nano: Just how small can you go?huggingface.co · 2025-10-31
  13. ArticleBuild a local AI co-pilot using IBM Granite Code, Ollama, and Continueallthingsopen.org · 2024-08-06
  14. ArticleIntel Xeon 6700P and 6500P Granite Rapids-SP for the Masses Initial Benchmarksservethehome.com · 2025-02-24
  15. Articlehttps://ibmgranite.mintlify.app/granite/docs/models/doclingibmgranite.mintlify.app · undated
  16. Articlegranite-docling-258M by ibm-granite | Mixpeek Model Hubmixpeek.com · undated
  17. ArticleIBM AI Releases Granite-Docling-258M: An Open-Source, Enterprise-Ready Document AI Model - MarkTechPostmarktechpost.com · 2025-09-18

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: 16 quantized builds and 22 Spaces.

  • As a service0no API provider listed
  • On own hardware16quantized builds
  • Built on2215 fine-tunes · 7 adapters · 0 merges
  • In apps22Hugging Face Spaces

Top Spaces running it: ibm-granite/granite-docling-258m-demo ♥ 282 · ibm-granite/granite-docling-258M-WebGPU ♥ 162 · jairwaal/image ♥ 16 · fmeres/granite-docling-demo ♥ 2

Will it fit?

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

Fits a 16 GB card at INT4

About 0.2 GB at INT4, 0.3 GB at FP8, 0.6 GB at BF16: weights only, so leave room for the KV cache.

16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 0.2 GB of weights (estimate; KV cache and activations extra)INT40.2 GBFP8: about 0.3 GB of weights (estimate; KV cache and activations extra)FP80.3 GBBF16: about 0.6 GB of weights (estimate; KV cache and activations extra)BF160.6 GB
16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 0.2 GB of weights (estimate; KV cache and activations extra)INT40.2 GBFP8: about 0.3 GB of weights (estimate; KV cache and activations extra)FP80.3 GBBF16: about 0.6 GB of weights (estimate; KV cache and activations extra)BF160.6 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.

9 dated articles and 30 repo commits so far

Teal is the weights (2025-05-19), amber triangles are 9 dated articles (each links to it), grey ticks are repo commits.

DecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2025-09-23: 1 repo commit2025-09-18: 3 repo commits2025-09-17: 21 repo commits2025-09-16: 5 repo commitsThe IBM Granite 4.1 family of models · 2026-04-30Granite 4.1: IBM's 8B Model Matching 32B MoE · 2026-04-30IBM's new Granite 4.2 models ride the wave of interest in local LLMs · 2026-08-26IBM Granite: A Family of Open Foundation Models for Code Intelligence · 2024-05-07IBM Granite 3.0: open enterprise models · 2024-10-21Open sourcing IBM's Granite code models · 2024-05-07IBM Granite 4.0 · 2025-10-02'Western Qwen': IBM Wows with Granite 4 LLM Launch and Hybrid Mamba/Transformer · 2025-10-03IBM Research – Granite Foundation Models Technical Report · 2023-11-29Weights 2025-05-19 · 2025-05-19Weights 2025-05-19now
DecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOctNovDecJanFebMarAprMayJunJulAugSepOct2025-09-23: 1 repo commit2025-09-18: 3 repo commits2025-09-17: 21 repo commits2025-09-16: 5 repo commitsThe IBM Granite 4.1 family of models · 2026-04-30Granite 4.1: IBM's 8B Model Matching 32B MoE · 2026-04-30IBM's new Granite 4.2 models ride the wave of interest in local LLMs · 2026-08-26IBM Granite: A Family of Open Foundation Models for Code Intelligence · 2024-05-07IBM Granite 3.0: open enterprise models · 2024-10-21Open sourcing IBM's Granite code models · 2024-05-07IBM Granite 4.0 · 2025-10-02'Western Qwen': IBM Wows with Granite 4 LLM Launch and Hybrid Mamba/Transformer · 2025-10-03IBM Research – Granite Foundation Models Technical Report · 2023-11-29Weights 2025-05-19 · 2025-05-19Weights 2025-05-19now

What people built on it

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

38 derivatives on the Hub, led by 16 quantized

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

16 quantized (Hugging Face model tree)15 fine-tunes (Hugging Face model tree)7 adapters (Hugging Face model tree)16 quantized15 fine-tunes7 adapters
16 quantized (Hugging Face model tree)15 fine-tunes (Hugging Face model tree)7 adapters (Hugging Face model tree)16 quantized15 fine-tunes7 adapters
GGUF ggml-org/granite-docling-258M-GGUF for llama.cpp, Ollama, LM Studio 51.3k downloads MLX ibm-granite/granite-docling-258M-mlx for Apple silicon 42.3k downloads

All variants →

Reference

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

What the card says

2 cited sentences

Summary generated, cited

Granite Docling is a 258M parameter multimodal image-text-to-text model designed for efficient document conversion.[1][2] The architecture is based on Idefics3 and is released under the Apache 2.0 license.[3][4]

  1. parameters human
  2. “multimodal Image-Text-to-Text model engineered for effici…”
  3. “builds upon the Idefics3 architecture”
  4. license

Generated 2026-10-07 15:04 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.

Use-case evidence

6 use cases

What it is good at 6 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

22 fields

Listed in LLMs

Facts

Pipeline tagimage-text-to-text
Librarytransformers
ArchitecturesIdefics3ForConditionalGeneration
Model typeidefics3
Dtypebfloat16
Params by dtypeBF16: 257,517,120
Size bandS
MoEno
Context length8,192 (config.json max_position_embeddings)
Languagesen
PapersarXiv:2501.17887, arXiv:2503.11576, arXiv:2305.03393
LicenseApache 2.0 (apache-2.0)
Hub snapshot2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · Idefics3ForConditionalGeneration · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: not in registry · 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 granite-docling-258M: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)0.6 GB · Edge (<4 GB)
FP8 (1 byte/param)0.3 GB · Edge (<4 GB)
INT4 (0.5 bytes/param)0.2 GB · Edge (<4 GB)

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

All variants

38 variants

Lineage

Original model Family: ibm-granite/granite 38 derivatives on the Hub

quantized: 16finetune: 15adapter: 7

Top variants (38 shown, ranked by downloads)

Relation
Quant
38 rows
Variant Relation Quant Params Band License Author Likes Downloads
ggml-org/granite-docling-258M-GGUF quantized GGUF — — apache-2.0 ggml-org 16 51.3k
ibm-granite/granite-docling-258M-mlx finetune mlx 315M S apache-2.0 ibm-granite 103 42.3k
onnx-community/granite-docling-258M-ONNX quantized — — — apache-2.0 onnx-community 7 33.7k
danchev/ibm-granite-docling-258M-GGUF quantized GGUF — — apache-2.0 danchev 4 5.8k
gguf-org/docling-gguf quantized gguf — — apache-2.0 gguf-org 6 5.3k
prithivMLmods/granite-docling-258M-f32-GGUF quantized GGUF — — apache-2.0 prithivMLmods 3 3.6k
docling-project/granite-docling-2stage-258m finetune — 258M S apache-2.0 docling-project 4 2.9k
philipp-zettl/granite-docling-258M-Q8_0-GGUF quantized Q8_0 — — apache-2.0 philipp-zettl 0 2.2k
infil00p/granite-docling-258M-GGUF quantized GGUF — — apache-2.0 infil00p 1 2.1k
SandLogicTechnologies/granite-docling-258M-GGUF quantized GGUF — — apache-2.0 SandLogicTechnologies 0 755
lamco-development/granite-docling-258M-onnx quantized — — — apache-2.0 lamco-development 7 337
padeoe/granite-docling-258M-GGUF quantized GGUF — — apache-2.0 padeoe 0 458
litert-community/granite-docling-258M quantized — — — apache-2.0 litert-community 0 370
tevfikk/granite-docling-258M-bitnet finetune — 258M S apache-2.0 tevfikk 1 238
AXONVERTEX-AI-RESEARCH/granite-docling-258M-Q8_0-GGUF quantized Q8_0 — — apache-2.0 AXONVERTEX-AI-RESEARCH 0 230
Kimhi/granite-docling-res-gate-lora adapter — — — apache-2.0 Kimhi 1 170
dongzi01/swarm adapter GGUF — — apache-2.0 dongzi01 0 160
GaborMadarasz/granite-docling-merged finetune — 258M S — GaborMadarasz 0 159
kizzlah/granite-docling-258M-mlx finetune mlx 315M S apache-2.0 kizzlah 0 119
pwsawbty/BodoWave finetune — — — apache-2.0 pwsawbty 2 0
furkankucuk/ResuVL-S finetune — 258M S cc-by-nc-4.0 furkankucuk 0 95
barflyman/granite-docling-258M-ONNX quantized — — — apache-2.0 barflyman 0 76
smdesai/granite-docling-258M-coreml quantized — — — apache-2.0 smdesai 0 63
textagent/granite-docling-258M-ONNX quantized — — — apache-2.0 textagent 0 60
Ak137/granite-docling-258M-finetuned finetune — — — — Ak137 1 0
DungND1107/IBM-granite-adapter-stage1 adapter — — — — DungND1107 0 46
khalidchawtany/ckb-granite_docling adapter — — — — khalidchawtany 0 33
DungND1107/ibm_finetuned_stage1 adapter — — — — DungND1107 0 31
JasonYANG170/granite-docling-258M-ONNX quantized — — — apache-2.0 JasonYANG170 0 21
Srinjoy99/tadpole-nano-258m finetune — 258M S apache-2.0 Srinjoy99 0 21
jitendrairor123/new-model finetune — — — mit jitendrairor123 0 0
holen777/test finetune — — — apache-2.0 holen777 0 0
Vane111/H finetune — — — bigscience-openrail-m Vane111 0 0
NexiriTech/Nexiri adapter — — — apache-2.0 NexiriTech 0 0
Berlinnad/VenTY adapter — — — apache-2.0 Berlinnad 0 0
wonwoostar/wonwoo finetune — — — apache-2.0 wonwoostar 0 0
hirsch-sw/slm_tables finetune — — — — hirsch-sw 0 0
smdesai/granite-docling-258M-coreai finetune — — — apache-2.0 smdesai 0 0

Repository changelog

30 commits
changed in the last 7 daysOther: 2026-10-24all changes

Repository changelog 30 commits

DateCommitSHA
2025-09-23 Add Troubleshooting section to README (#23) 982fe3b
2025-09-18 Update README.md 377322a
2025-09-18 token_alignment (#13) 97b0a53
2025-09-18 Adding `transformers` tag to the library field (#11) 70e26c5
2025-09-17 Update README.md a1ff0fe
2025-09-17 Update README.md 5289190
2025-09-17 Update README.md da240aa
2025-09-17 Update README.md 0281789
2025-09-17 Update README.md 9e1ee78
2025-09-17 Update README.md (#9) a709291
2025-09-17 Update README.md 2e257e0
2025-09-17 Update README.md (#8) b0d56bd
2025-09-17 Update README.md 3299088
2025-09-17 Upload granite_docling_split_page.png 3e1f4b0
2025-09-17 Update README.md 906d75c
2025-09-17 Update README.md 46d5666
2025-09-17 Update README.md 595d39b
2025-09-17 Update README.md 24a14e1
2025-09-17 Update README.md 65f280a
2025-09-17 Update README.md f0856a8
2025-09-17 Update README.md (#7) 8903fa1
2025-09-17 Update README.md 9a7beaf
2025-09-17 Update README.md (#6) a0903a5
2025-09-17 Upload assets/new_arxiv.png with huggingface_hub 3beb7d8
2025-09-17 Update README.md (#5) 40d7a97
2025-09-16 Update README.md 5ead9c9
2025-09-16 Update README.md 7e97417
2025-09-16 Update README.md cbc889a
2025-09-16 Update README.md cd1a8df
2025-09-16 Update README.md 6d56bf5

Articles and coverage

19 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.

Official6

The IBM Granite 4.1 family of models
research.ibm.com2026-04-30officialauthorityabout the granite familyHN 88 pts · 16 comments1,677 words
IBM released the Granite 4.1 family of models under an Apache 2.0 license, offering language models in 3B, 8B, and 30B parameter sizes alongside vision, speech, embedding, and guardian capabilities. These language models are trained on approximately 15 trillion tokens and extend context length to as much as 512K tokens, while the 97M-parameter embedding model supports retrieval across more than 200 languages. Developers can access these models on watsonx, Hugging Face, and via optimized inference runtimes including vLLM, SGLang, and llama.cpp.written, reviewed, cited
  • Granite Speech 4.1 2B NAR generates entire sequences at once instead of one token at a time.
  • The Granite 4.1 8B instruct model matches or outperforms the Granite 4.0 32B Mixture-of-Experts model.
  • Granite Guardian 4.1 is a direct replacement for Granite Guardian 3.3 8B and was fine-tuned on top of Granite 4.1 8B.
IBM Granite-Docling: End-to-end document understanding
ibm.comofficialauthorityabout the granite family1,433 words
IBM releases Granite-Docling-258M, an ultra-compact open-source vision-language model with 258M parameters available under an Apache 2.0 license for converting documents to machine-readable formats. The model utilizes the DocTags markup format to preserve layout, tables, and equations, offering experimental multilingual capabilities for Arabic, Chinese, and Japanese while future iterations target sizes of approximately 512M and 900M parameters.written, reviewed, cited
  • Granite-Docling replaces the SmolLM-2 language backbone used for SmolDocling with a Granite 3-based architecture.
  • The model addresses instabilities in its predecessor by filtering datasets to remove samples with inconsistent or missing annotations.
  • Experimental multilingual capabilities for Arabic, Chinese, and Japanese have not yet been validated for enterprise-ready performance or stability.
IBM Granite 3.0: open enterprise models
ibm.com2024-10-21officialauthorityabout the granite familyHN 39 pts · 6 comments1,962 words
IBM releases Granite 3.0, featuring dense models like Granite 3.0 8B Instruct and Granite 3.0 2B Instruct alongside mixture of experts (MoE) variants for minimum latency. All models are released under the permissive Apache 2.0 license and available today for commercial use on the IBM watsonx platform. The collection includes LLM-based input-output guardrail models and a speculative decoder for increased inference speed and efficiency.written, reviewed, cited
  • IBM provides an uncapped indemnity for third party IP claims against IBM-developed models.
  • The IBM Power scheduler adjusts the learning rate based on the number of tokens and batch size based on a power-law equation.
Open sourcing IBM's Granite code models
research.ibm.com2024-05-07officialauthorityabout the granite familyHN 18 pts · 4 comments1,880 words
IBM Granite 4.0
ibm.com2025-10-02officialauthorityabout the granite familyHN 10 pts · 3 comments2,656 words
IBM launches Granite 4.0, a new family of open language models featuring a hybrid Mamba/transformer architecture designed for enterprise efficiency. Released under an Apache 2.0 license, these models are the first open models to receive ISO 42001 certification and offer over 70% reduction in RAM for long inputs. The hybrid architecture combines Mamba-2 layers with transformer attention, and IBM offers an uncapped indemnity for third party IP claims alongside a bug bounty program for security flaws.written, reviewed, cited
  • Granite 4.0-H-Small exceeds all open weight models on IFEval except Llama 4 Maverick, a 402B-parameter model over 12 times its size.
  • Granite 4.0 models are compatible with AMD Instinct MI-300X GPUs, enabling even further reduction of their memory footprint.
IBM Research – Granite Foundation Models Technical Report
ibm.com2023-11-29officialauthorityabout the granite familyHN 4 pts · 2 comments

Analysis4

Granite 4.1: IBM's 8B Model Matching 32B MoE
firethering.com2026-04-30HN front pageabout the granite familyHN 328 pts · 209 comments1,865 words
IBM released Granite 4.1, a family of open source models trained on 15 trillion tokens with a rigorous data quality pipeline. The 8B dense model outperforms the previous 32B MoE variant on benchmarks like ArenaHard and GSM8K. Development involved four sequential RL stages and resulted in 4.1 million curated samples.written, reviewed, cited
  • The 3B model extends context only to 128K tokens, not the 512K available to larger variants.
  • Phase 5 extended the context window eventually to 512K tokens for the 8B and 30B models.
  • Stage three was a short identity and knowledge calibration run about 40 training steps to stabilize representation.
IBM's new Granite 4.2 models ride the wave of interest in local LLMs
arstechnica.com2026-08-26about the granite familyHN 4 pts · 0 comments584 words
IBM has released the Granite 4.2 family of open-weight large language models designed for local deployment. The 8B and 30B variants undergo an agentic reinforcement-learning block to enable capabilities like using the terminal, searching the web, or using external tools, while the 3B model supports tools without specialized training. This reasoning-focused release offers a native 128,000-token context window but often incurs slower response times and higher compute demands compared to competitors like Nvidia's Nemotron.written, reviewed, cited
IBM Granite: A Family of Open Foundation Models for Code Intelligence
github.com2024-05-07authorityHN front pageabout the granite familyHN 252 pts · 74 comments1,032 words
IBM introduces the Granite series of decoder-only code models trained on 116 programming languages. All models are released under an Apache 2.0 license for research and commercial use. Base models are trained on 3-4T tokens of code data and natural language datasets. The family includes base and instruct variants available in 3B, 8B, 20B, and 34B parameter sizes.written, reviewed, cited
  • The 34B model is trained on 1.4T tokens after depth upscaling done on the 1.6T checkpoint of 20B model.
  • We apply a HAP content filter that reduces models' likelihood of generating hateful, abusive, or profane language.
  • We also scan all datasets using ClamAV to identify and remove instances of malware in the source code.
'Western Qwen': IBM Wows with Granite 4 LLM Launch and Hybrid Mamba/Transformer
venturebeat.com2025-10-03about the granite familyHN 83 pts · 25 comments
All coverage tracked for this model (19) — append-only, nothing is ever removed
TitleSourcePublishedFirst seenSignals
IBM Granite-Docling: End-to-end document understandingibm.com official—2026-10-07exa
https://ibmgranite.mintlify.app/granite/docs/models/doclingibmgranite.mintlify.app—2026-10-07exa
granite-docling-258M by ibm-granite | Mixpeek Model Hubmixpeek.com—2026-10-07exa
IBM's new Granite 4.2 models ride the wave of interest in local LLMsarstechnica.com2026-08-262026-10-07hn
Fun Local LLM Comparisons with Gemma, Granite, and Qwenekorbia.com2026-05-282026-10-07hn
Granite 4.1: IBM's 8B Model Matching 32B MoEfirethering.com2026-04-302026-10-07hn
The IBM Granite 4.1 family of modelsresearch.ibm.com official2026-04-302026-10-07hn
Show HN: A local-first financial auditor using IBM Granite, MCP, and SQLitegithub.com2026-01-012026-10-07hn
Granite 4.0 Nano: Just how small can you go?huggingface.co2025-10-312026-10-07hn
'Western Qwen': IBM Wows with Granite 4 LLM Launch and Hybrid Mamba/Transformerventurebeat.com2025-10-032026-10-07hn
IBM Granite 4.0ibm.com official2025-10-022026-10-07hn
IBM AI Releases Granite-Docling-258M: An Open-Source, Enterprise-Ready Document AI Model - MarkTechPostmarktechpost.com2025-09-182026-10-07exa
Intel Xeon 6700P and 6500P Granite Rapids-SP for the Masses Initial Benchmarksservethehome.com2025-02-242026-10-07hn
IBM Granite 3.0: open enterprise modelsibm.com official2024-10-212026-10-07hn
Build a local AI co-pilot using IBM Granite Code, Ollama, and Continueallthingsopen.org2024-08-062026-10-07hn
IBM open-sources its Granite AI models – and they mean businesszdnet.com2024-05-132026-10-07hn
Open sourcing IBM's Granite code modelsresearch.ibm.com official2024-05-072026-10-07hn
IBM Granite: A Family of Open Foundation Models for Code Intelligencegithub.com2024-05-072026-10-07hn
IBM Research – Granite Foundation Models Technical Reportibm.com official2023-11-292026-10-07hn

Papers

3 papers

Papers 3

arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).

Similar models

3 models

Similar models

  • Florence-2-base · 232M
    same size band (S), same task (image-text-to-text), permissive license
  • Florence-2-base-ft · 232M
    same size band (S), same task (image-text-to-text), permissive license
  • LFM2.5-VL-450M · 449M
    same size band (S), same task (image-text-to-text)

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:36 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:04 UTC · Last verified 2026-10-25

Follow granite-docling-258M

Watch this model

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