Model release

Zamba2-VL-7B

Zamba2-VL-7B is an open-weight model release in the Zamba family from Zyphra1 2, version Zamba2-VL-7B, released Jun 2, 2026. Its weights can be downloaded by the public1 under the Apache License 2.01 2.

The largest of the three Zamba2-VL vision-language models Zyphra released in June 2026, with about 8B parameters. It pairs the Zamba2-7B hybrid state-space language model with the Qwen2.5-VL vision encoder and handles single- and multi-image understanding and grounding.2 1

Model hub: Model card (Hugging Face) (external site: huggingface.co)
Model hub: Zamba2-VL collection (Hugging Face) (external site: huggingface.co)
Release notes: Zamba2-VL announcement (external site: zyphra.com)
Paper: Zamba2-VL technical report (arXiv 2606.00390) (external site: arxiv.org)
Repository: Inference code (Zyphra Transformers fork, zamba2-vl branch) (external site: github.com)

On AI on Hugging Face: Zyphra/Zamba2-VL-7B

Last reviewedSep 29, 2026Entry updated Sep 29, 2026Documented release Jun 2, 2026

Availability and license

Overall availability

Public

Downloadable from Hugging Face without gating.1

Availability fact review: not separately recorded

Availability is separate from permission: read the license before using or redistributing.

Apache License 2.0 (external site: huggingface.co)1 2
Applies to: Weights · SPDX Apache-2.0 · on the rubric’s OSI-approved software-license list · fact-level rights review not recorded
Plain-language guide to the Apache License 2.0
Apache License 2.0 (Zyphra Transformers fork) (external site: raw.githubusercontent.com)4
Applies to: Code · SPDX Apache-2.0 · on the rubric’s OSI-approved software-license list · fact-level rights review not recorded
Plain-language guide to the Apache License 2.0

Component reuse rights

A readable or downloadable component is not automatically reusable. These indicators concern recorded license evidence, not system certification.

weights: Unknown — no complete fact-level rights review
code: Unknown — no complete fact-level rights review
data: Unknown — no complete fact-level rights review
documentation: Unknown — no complete fact-level rights review

No complete system-rights review is recorded for this release.

Model-disclosure tier

Computed from the checklist below using USASI rubric v0.2. An editorial category, not a certification.

Open-weight

The model parameters for this release can be downloaded by the public. License terms may still restrict use, redistribution, or commercial use.

How tiers are computed

Public materials checklist

Items for a model under USASI rubric v0.2. Unknown means unassessed or insufficient evidence.

Public materials checklist for Zamba2-VL-7B
Item	Status	Notes and evidence
WeightsCan the general public download the model parameters for this release?	Public	1
Inference codeIs code for running the model published?	Public	Inference uses the zamba2-vl branch of Zyphra's Transformers fork (based on Transformers 4.57.1), qwen-vl-utils, flash_attn, and Zyphra's forks of mamba-ssm and causal-conv1d, per the model card.1 2
Training codeIs the code used to train the model published?	Unknown	The technical report points to the checkpoints and inference code; it does not mention a training code release.3
Training-data informationDoes the information cover provenance, scope, acquisition, selection, labeling, processing, and where data or alternatives can be obtained? Access alone does not establish completeness.	Unknown	Not assessed.
Training-data accessCan the training data be obtained? This is independent of information completeness and reuse rights; original unshareable data need not be downloadable.	Unknown	Not assessed.
Complete training pipelineIs the complete base-training and preprocessing pipeline published, including configuration? Fine-tuning code or an inference SDK alone is insufficient.	Unknown	Not assessed.
Legacy data assessment (v0.1)Historical assessment combining download access and disclosure. Preserved for traceability; excluded from the v0.2 tier calculation. See the new separate assessments above.	Partial	The technical report names the public datasets used in its training stages, for example LLaVA-ReCap-558K, FineVision, PixMo, The Cauldron, and DocMatix, and says document and OCR data were upsampled.3
Training recipeAre the training configuration and procedure documented in enough detail to follow?	Partial	The report describes three stages: alignment with only the MLP adapter trained (vision encoder and language backbone frozen), then joint pretraining and supervised fine-tuning of the full model.3
Evaluation materialsPublic = evaluation code or prompts that let others re-run the evaluations are published. Partial = results only.	Partial	The model card reports scores measured with Zyphra's own evaluation harness based on VLMEvalKit; the harness itself is not linked.1

What it is useful for

The model card describes it as a generalist model for on-device applications and reports results on visual question answering, OCR, chart and document understanding, and multimodal reasoning benchmarks. Zyphra's announcement says the hybrid backbone keeps a fixed-size recurrent state rather than a growing key-value cache, which lowers time to first token on long inputs.1 2

Run and use notes

Documented facts only. No hardware or performance claims are made without a cited source and stated assumptions.

The model card says the model can run without the optimized Mamba2 kernels but that this is not recommended because latency and memory use are significantly higher. Its example loads the model in bfloat16 on a CUDA device with FlashAttention 2.1

Organization context

Zyphra

Provenance and derivatives

Built by Zyphra on its Zamba2-7B language model, with the vision transformer from Qwen2.5-VL as the vision backbone; Qwen is a model family built by Alibaba Cloud. The model uses the Mistral v0.1 tokenizer.1 2 6

Derived from: Zamba2-7B — Language backbone.
Derived from: Qwen2.5-VL vision encoder (external site: huggingface.co) — Vision backbone, developed by Alibaba Cloud's Qwen team. The model card and technical report do not say which Qwen2.5-VL checkpoint supplied the encoder.

Other releases in the Zamba family

Zamba2-7BModel-disclosure tier (USASI rubric v0.2): Open-weight

Zamba family overview

What this catalog does not know

Unknown means the sources reviewed for this record do not document it. It is not evidence that something does not exist.

Training code: unknown.
Training-data information: unknown.
Training-data access: unknown.
Complete training pipeline: unknown.

Have a primary source? How to report a correction.

U.S. eligibility

Project eligibility rests on documented governing or maintaining entities, not on contributors.

Eligible · basis: U.S. headquarters

Developed and published by Zyphra. The technical report lists the authors' affiliation as Zyphra, San Francisco, CA, and Zyphra's terms of use name Zyphra Technologies, Inc. at a San Francisco address. The vision encoder comes from Alibaba Cloud's Qwen2.5-VL; that component is not thereby U.S.-developed, and eligibility here rests on Zyphra as the developer of this model.3 5 1 6

Assessed Sep 29, 2026

Sources

1.
Zyphra/Zamba2-VL-7B model card (external site: huggingface.co)
Zyphra · Model card · accessed Sep 29, 2026
2.
Zamba2-VL: Hybrid SSM Vision-Language Models (external site: zyphra.com)
Zyphra · Announcement · published Jun 2, 2026 · accessed Sep 29, 2026
3.
Zamba2-VL Technical Report (arXiv 2606.00390v1) (external site: arxiv.org)
Zyphra · Paper · published May 29, 2026 · accessed Sep 29, 2026
4.
Zyphra/transformers LICENSE (zamba2-vl branch) (external site: raw.githubusercontent.com)
Hugging Face (fork maintained by Zyphra) · License · accessed Sep 29, 2026
5.
Terms of Use | Zyphra (external site: zyphra.com)
Zyphra Technologies, Inc. · Official page · published May 4, 2026 · accessed Sep 29, 2026
6.
Qwen on Hugging Face (external site: huggingface.co)
Qwen (Alibaba Cloud) · Repository · accessed Sep 29, 2026

Email a correction Report a correction Edit this entry

This listing is not an endorsement, a safety assessment, or a federal approval.

Cite this page

APA style

USASI. (2026, September 29). Zamba2-VL-7B. United States of America Superintelligence. https://unitedstatesofamericasuperintelligence.com/open/zamba2-vl-7b/

BibTeX

@misc{usasi_zamba2_vl_7b,
  author = {{USASI}},
  title = {Zamba2-VL-7B},
  year = {2026},
  month = sep,
  howpublished = {\url{https://unitedstatesofamericasuperintelligence.com/open/zamba2-vl-7b/}},
  note = {United States of America Superintelligence. Last updated 2026-09-29}
}

The date is when this page was last updated. Add the date you read it if your style needs one.

For maintainers: a “Listed in USASI” badge

Links to this page. It says the entry is listed, not endorsed or rated. About the badge

Markdown (for a README)

[![Listed in the USASI catalog](https://unitedstatesofamericasuperintelligence.com/badges/listed.svg)](https://unitedstatesofamericasuperintelligence.com/open/zamba2-vl-7b/)

HTML

<a href="https://unitedstatesofamericasuperintelligence.com/open/zamba2-vl-7b/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/listed.svg" alt="Listed in the USASI catalog" height="20"></a>