selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Alternatives to d1-2B

120 open-weight models that do the same job as d1-2B (1.9B vision), ranked by how close they are in task and size, then by popularity and how recent they are. Licence and VRAM are shown for each.

the original
LiquidAI logo

d1-2B

LiquidAI·Vision

≈2.0 GB

d1-2B is a decision model for text, pictures, and video, built on LFM3-VL-2B. It returns calibrated answers in one forward pass, without generating tokens.

1.9B125K ctx lfm1.0
16.2KOct 2026ollama
--vram
any
--gpus
--quant
  1. microsoft logo

    Mage-VL

    microsoft·Vision

    ≈4.4 GB

    Mage-VL is a codec-native, proactive-streaming multimodal foundation model for image and video understanding, whose visual encoder is trained entirely from scratch at a compact 4B scale.

    4.7B256K ctx Apache 2.0
    7KJul 2026gguf
    • Same category
    • Larger (4.7B vs 1.9B)
    • Commercial use OK
    vs
  2. openbmb logo

    MiniCPM-V-4

    openbmb·Vision

    ≈3.2 GB

    A GPT-4V Level MLLM for Single Image, Multi Image and Video on Your Phone

    4.1B32K ctx Apache 2.0
    29.6KJul 2025gguf
    • Same category
    • Larger (4.1B vs 1.9B)
    • Commercial use OK
    vs
  3. nvidia logo

    Nemotron-3.5-Parse

    nvidia·Vision

    ≈1.2 GB

    Nemotron 3.5 Parse is a document image parser that produces structured, machine-readable text with semantic layout classes, bounding boxes, and reading order. It combines a C-RADIOv4-SO400M vision encoder with a 10-layer mBART decoder that consumes the visual sequence as a causal prefix.

    631M16K ctx Custom licence
    27.1KSept 2026gguf
    • Same category
    • Smaller (631M vs 1.9B)
    • Commercial use OK
    vs
  4. maritaca-ai logo

    S5-nano

    maritaca-ai·Vision

    ≈4.3 GB

    S5-nano is a 4B-parameter model from Maritaca AI. It has two modes:

    4.7B256K ctx Apache 2.0
    2.2KNov 2026gguf
    • Same category
    • Larger (4.7B vs 1.9B)
    • Commercial use OK
    • Newer release
    vs
  5. ibm-granite logo

    granite-4.0-3b-vision

    ibm-granite·Vision

    ≈3.5 GB

    Model Summary: Granite-4.0-3B-Vision is a vision-language model (VLM) designed for enterprise-grade document data extraction. It focuses on specialized, complex extraction tasks that ultracompact models often struggle with:

    4B128K ctx Apache 2.0
    3.2KMar 2026gguf
    • Same category
    • Larger (4B vs 1.9B)
    • Commercial use OK
    vs
  6. allenai logo

    Molmo2-4B

    allenai·Vision

    ≈4.5 GB

    Molmo2 is a family of open vision-language models developed by the Allen Institute for AI (Ai2) that support image, video and multi-image understanding and grounding.

    4.9B36K ctx Apache 2.0
    32.8KDec 2025gguf
    • Same category
    • Larger (4.9B vs 1.9B)
    • Commercial use OK
    vs
  7. google logo

    t5gemma-2-270m-270m

    google·Vision

    ≈1.1 GB

    T5Gemma is a family of lightweight yet powerful encoder-decoder research models from Google, built by adapting pretrained decoder-only Gemma models into encoder-decoder ones.

    786M32K ctx Gemma Terms of Use
    33.2KOct 2025
    • Same category
    • Smaller (786M vs 1.9B)
    vs
  8. ibm-granite logo

    granite-vision-3.2-2b

    ibm-granite·Vision

    ≈2.9 GB

    Model Summary: granite-vision-3.2-2b is a compact and efficient vision-language model, specifically designed for visual document understanding, enabling automated content extraction from tables, charts, infographics, plots, diagrams, and more.

    3B128K ctx Apache 2.0
    3.6KFeb 2025gguf
    • Same category
    • Similar size (3B vs 1.9B)
    • Commercial use OK
    vs
  9. microsoft logo

    kosmos-2-patch14-224

    microsoft·Vision

    ≈1.9 GB

    This Hub repository contains a HuggingFace's transformers implementation of the original Kosmos-2 model from Microsoft.

    1.7BMIT
    7.7KOct 2023
    • Same category
    • Similar size (1.7B vs 1.9B)
    • Commercial use OK
    vs
  10. microsoft logo

    Fara1.5-4B

    microsoft·Vision

    ≈4.2 GB

    Fara1.5-4B is a multimodal computer use agent (CUA) for web browsers, from Microsoft Research AI Frontiers. It observes the browser through screenshots and acts on the user's behalf by emitting structured tool calls — click, type, scroll, visit URL, web search, and so on — to complete tasks…

    4.5B256K ctx MIT
    2.8KJul 2026gguf
    • Same category
    • Larger (4.5B vs 1.9B)
    • Commercial use OK
    vs
  11. microsoft logo

    Phi-3.5-vision-instruct

    microsoft·Vision

    ≈6.0 GB

    Phi-3.5-vision is a lightweight, state-of-the-art open multimodal model built upon datasets which include - synthetic data and filtered publicly available websites - with a focus on very high-quality, reasoning dense data both on text and vision.

    4.1B128K ctx MIT
    241.1KAug 2024
    • Same category
    • Larger (4.1B vs 1.9B)
    • Commercial use OK
    vs
  12. Qwen logo

    Qwen3.5-9B

    Qwen·Vision

    ≈7.2 GB

    Over recent months, we have intensified our focus on developing foundation models that deliver exceptional utility and performance.

    9.7B256K ctx Apache 2.0
    6.4MFeb 2026ollama
    • Same category
    • Larger (9.7B vs 1.9B)
    • Commercial use OK
    • Widely used
    vs
  13. ibm-granite logo≈2.9 GB

    Model Summary: granite-vision-3.1-2b-preview is a compact and efficient vision-language model, specifically designed for visual document understanding, enabling automated content extraction from tables, charts, infographics, plots, diagrams, and more.

    3B16K ctx Apache 2.0
    2.7KJan 2025gguf
    • Same category
    • Similar size (3B vs 1.9B)
    • Commercial use OK
    vs
  14. llava-hf logo≈1.1 GB

    Check out also the Google Colab demo to run Llava on a free-tier Google Colab instance:

    894MApache 2.0
    189.2KAug 2024
    • Same category
    • Smaller (894M vs 1.9B)
    • Commercial use OK
    vs
  15. nvidia logo

    DAM-3B-Video

    nvidia·Vision

    ≈2.9 GB

    Describe Anything Model 3B Video (DAM-3B-Video) takes inputs of user-specified regions in the form of points/boxes/scribbles/masks within images/videos, and generates detailed localized descriptions of images/videos.

    3Bnvidia-non-commercial-license
    1.5KApr 2025
    • Same category
    • Similar size (3B vs 1.9B)
    vs
  16. inclusionAI logo

    ZwZ2-4B

    inclusionAI·Vision

    ≈4.2 GB

    ZwZ2-4B is the Qwen3.5-4B-based model in Zooming-without-Zooming-v2 (ZwZ2). It learns to connect visual evidence across regions through comparison, aggregation, matching, and multi-step reasoning. Training uses full images; inference requires no zooming tools.

    4.5B256K ctx Unknown licence
    1.1KOct 2026gguf
    • Same category
    • Larger (4.5B vs 1.9B)
    vs
  17. google logo

    paligemma2-3b-pt-224

    google·Vision

    ≈3.2 GB

    Transformers PaliGemma 2 3B weights, pre-trained with 224*224 input images and 128 token input/output text sequences. The model is available the bfloat16 format for fine-tuning.

    3BGemma Terms of Use
    4.8KNov 2024
    • Same category
    • Similar size (3B vs 1.9B)
    vs
  18. google logo

    paligemma2-3b-pt-448

    google·Vision

    ≈3.2 GB

    Transformers PaliGemma 2 3B weights, pre-trained with 448*448 input images and 512 token input/output text sequences. The model is available the bfloat16 format for fine-tuning.

    3BGemma Terms of Use
    5.1KNov 2024gguf
    • Same category
    • Similar size (3B vs 1.9B)
    vs
  19. Qwen logo≈6.8 GB

    Meet Qwen3-VL — the most powerful vision-language model in the Qwen series to date.

    8.8B256K ctx Apache 2.0
    6.2MOct 2025ollama
    • Same category
    • Larger (8.8B vs 1.9B)
    • Commercial use OK
    • Widely used
    vs
  20. BAAI logo≈2.0 GB

    The Aquila-VL-2B model is a vision-language model (VLM) trained based on the LLava-one-vision framework. The Qwen2.5-1.5B-instruct model is chose as the LLM, while siglip-so400m-patch14-384 is utilized as the vision tower.

    2.2B32K ctx Apache 2.0
    241Oct 2024
    • Same category
    • Similar size (2.2B vs 1.9B)
    • Commercial use OK
    vs
  21. nvidia logo

    NV-Reason-CT

    nvidia·Vision

    ≈4.7 GB

    NV-Reason-CT is a 3D vision-language model (VLM) for CT image analysis. It combines a native 3D vision encoder (3D ViT) with a language model and is designed for radiology report generation, general question answering, and multi-step reasoning across chest and abdominal CT volumes.

    5.3B256K ctx Custom licence
    4.3KSept 2026
    • Same category
    • Larger (5.3B vs 1.9B)
    • Commercial use OK
    vs
  22. Efficient-Large-Model logo

    VILA1.5-3b

    Efficient-Large-Model·Vision

    ≈2.9 GB

    Model type: VILA is a visual language model (VLM) pretrained with interleaved image-text data at scale, enabling multi-image VLM. VILA is deployable on the edge, including Jetson Orin and laptop by AWQ 4bit quantization through TinyChat framework.

    3BCC BY-NC 4.0
    16.2KApr 2024
    • Same category
    • Similar size (3B vs 1.9B)
    vs
  23. internlm logo

    CapRL-3B

    internlm·Vision

    ≈3.0 GB

    We are excited to release the CapRL 2.0 series: CapRL-Qwen3VL-2B and CapRL-Qwen3VL-4B. These models feature fewer parameters while delivering even more powerful captioning performance. Notably, CapRL-Qwen3VL-2B outperforms both CapRL-Qwen2.5VL-3B and Qwen2.5VL-72B in captioning tasks.

    3.8B125K ctx Apache 2.0
    2.4KSept 2025gguf
    • Same category
    • Similar size (3.8B vs 1.9B)
    • Commercial use OK
    vs
  24. Salesforce logo

    blip2-opt-2.7b

    Salesforce·Vision

    ≈5.2 GB

    BLIP-2 model, leveraging OPT-2.7b (a large language model with 2.7 billion parameters). It was introduced in the paper BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models by Li et al. and first released in this repository.

    3.7BMIT
    386.4KFeb 2023
    • Same category
    • Similar size (3.7B vs 1.9B)
    • Commercial use OK
    vs