Specification

Release date: 14 Dec 2025
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Openness: open-weights
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
License: apache-2.0
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Architecture: Molmo2ForConditionalGeneration
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Parameters: 8.66B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
Languages: en
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Base model: Qwen/Qwen3-8B, google/siglip-so400m-patch14-384
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
File size: 34.6 GB
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Hugging Face repo: allenai/Molmo2-8B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
Pipeline tag: image-text-to-text
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
Model card: huggingface.co/allenai/Molmo2-8B
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
Downloads: 94,605
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 54 min agomedium
Likes: 193
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 54 min agomedium
Datasets: allenai/Molmo2-Cap, allenai/Molmo2-VideoCapQA, allenai/Molmo2-VideoSubtitleQA, allenai/Molmo2-AskModelAnything, allenai/Molmo2-VideoPoint, allenai/Molmo2-VideoTrack, allenai/Molmo2-MultiImageQA, allenai/Molmo2-SynMultiImageQA, allenai/Molmo2-MultiImagePoint
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Gated: No
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Last modified: 2026-01-23T01:02:55+00:00
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 1 h agomedium
Library name: transformers
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Downloads all time: 3,183,400
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 54 min agomedium
Model type: molmo2
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Tags: transformers, safetensors, molmo2, image-text-to-text, multimodal, olmo, molmo, custom_code, en
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium
Weights dtype: F32
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium

Each value shows its source, tier and observation time. Conflicting claims are kept side by side and flagged — never averaged. How AI Atlas records facts →

Provenance

Attributed facts

21

Source tiers

T221

Freshest observation

54 min ago

Conflicts

None

Relations

Derived from: siglip-so400m-patch14-384
Developed by: Allen Institute for AI
Fine-tuned from: Qwen3 8B
Trained on: allenai/Molmo2-SynMultiImageQA, allenai/Molmo2-MultiImagePoint, allenai/Molmo2-VideoTrack, allenai/Molmo2-Cap, allenai/Molmo2-VideoCapQA, allenai/Molmo2-VideoSubtitleQA, allenai/Molmo2-AskModelAnything, allenai/Molmo2-MultiImageQA +1(9 total)

Identity

hf_repo: allenai/Molmo2-8B

Also known as: allenai/Molmo2-8B, Molmo2-8B

slug molmo2-8b

ModelMolmo2-4BAllen Institute for AI
ModelOlmo-3-1025-7BAllen Institute for AI
ModelOLMoE-1B-7B-0924Allen Institute for AI
ModelwildguardAllen Institute for AI
ModelOLMoE-1B-7B-0125-InstructAllen Institute for AI
ModelolmOCR-2-7B-1025Allen Institute for AI
Modelallenai/olmOCR-2-7B-1025-FP8Allen Institute for AI
ModelOlmo-3-7B-InstructAllen Institute for AI

Compare

Open compare →

Modalities

Modalities unavailable.

Capabilities

Tool calling
Unavailable
Structured output
Unavailable
Reasoning
Unavailable
Vision
Unavailable
Audio
Unavailable
Fine-tuning available
Unavailable

No capability flags have been observed from a source yet — we do not infer them.

Languages: en
Source:Hugging Face Hub (public pages, model cards, papers)T2observed 56 min agomedium

No benchmark results recorded

Results appear when a tier 1–3 source publishes them; we never copy scores without a source.

Current prices

No current prices recorded

Prices appear when a provider publishes a public pricing page we crawl.

Price history

Estimated

Memory need = bytes per parameter (4-bit ≈ 0.5 × 1.15 overhead, 8-bit 1.0, fp16 2.0) + a KV-cache allowance. Not a measurement.

Estimated hardware fit
Hardware	Quantization	Memory	Est. need	Fits
Apple M3 Ultra	4bit	—	5.5 GB	Yes
AMD Instinct MI325X	4bit	256 GB	5.5 GB	Yes
AMD Instinct MI300X	4bit	192 GB	5.5 GB	Yes
Apple M2 Ultra	4bit	—	5.5 GB	Yes
NVIDIA B200	4bit	180 GB	5.5 GB	Yes
NVIDIA H200	4bit	141 GB	5.5 GB	Yes
Apple M1 Ultra	4bit	—	5.5 GB	Yes
Apple M3 Max	4bit	—	5.5 GB	Yes
Apple M4 Max	4bit	—	5.5 GB	Yes
NVIDIA DGX Spark	4bit	128 GB	5.5 GB	Yes
Apple M2 Max	4bit	—	5.5 GB	Yes
NVIDIA A100 80GB	4bit	80 GB	5.5 GB	Yes
NVIDIA H100 SXM	4bit	80 GB	5.5 GB	Yes
Apple M1 Max	4bit	—	5.5 GB	Yes
Apple M4 Pro	4bit	—	5.5 GB	Yes
Apple M3 Pro	4bit	—	5.5 GB	Yes
Apple M1 Pro	4bit	—	5.5 GB	Yes
Apple M2 Pro	4bit	—	5.5 GB	Yes
Apple M4	4bit	—	5.5 GB	Yes
NVIDIA GeForce RTX 5090	4bit	32 GB	5.5 GB	Yes
Apple M2	4bit	—	5.5 GB	Yes
Apple M3	4bit	—	5.5 GB	Yes
NVIDIA GeForce RTX 3090	4bit	24 GB	5.5 GB	Yes
NVIDIA GeForce RTX 4090	4bit	24 GB	5.5 GB	Yes
Apple M1	4bit	—	5.5 GB	Yes

Ancestors 3

This model

Molmo2-8B

8.66B params

Quantizations 0

None recorded.

Descendants 0

None recorded.

Papers 0

No papers linked yet.

Repositories 0

No repositories linked yet.

As of

Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.

Claim history · Downloads

1 claims · 1 propertiesShow all properties

Downloadsmetric.downloads1

Claim history for Downloads
Value	Valid from → to	Status	Source	Confidence	Extractor
94,605	11 Sept 2026 21:10 UTC → current	current	Hugging Face Hub (public pages, model cards, papers)T2	medium	deterministic

Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →

11 Sept 2026
New modelModelMolmo2-8BAllen Institute for AI
New model: Molmo2-8B (Allen Institute for AI)
huggingface
huggingface.co

Full timeline →

Source documents
Source	Document	Type	Tier	Last observed	Snapshots
Hugging Face Hub (public pages, model cards, papers)	huggingface.co/allenai/Molmo2-8B	model_page	T2· Quality secondary	54 min ago	2
Hugging Face Hub (public pages, model cards, papers)	huggingface.co/models?author=allenai&p=0&sort=downloads	listing	T2· Quality secondary	57 min ago	2

Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.