BL

BlazerApex Vision V4 MM

BlazerapexGGUF by Davizig10jojoUpdated 2026-10-09
Download on Hugging Face
Parameters1.88B
Context256K
Layers24
Architectureqwen35
Smallest quant1.2 GB
Quick answer

BlazerApex Vision V4 MM needs about 2.1 GB of memory at its main file (a 1.2 GB file) with an 8K context window, so it fits on a 4 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 1.6 GB.

BlazerApex Vision V4 MM – multimodal 1.88B parameter model

BlazerApex Vision V4 MM is a 1.88B parameter multimodal model based on Qwen3.5. It supports image‑to‑text tasks, reasoning, tools, and Portuguese Brazilian language. The model is experimental and uncensored.

Good at

  • Image description in Portuguese
  • Step‑by‑step reasoning
  • Python code analysis

Limits

  • Experimental, uncensored
  • No vision support without mmproj file
  • Requires 8k context for full memory usage

Best for

  • Developers building Portuguese chatbots
  • Researchers testing multimodal reasoning
  • Educational tools for code explanation

Who it’s for. The model is intended for developers and researchers who need a Brazilian Portuguese multimodal assistant with reasoning and coding capabilities. It is not recommended for production use without prior evaluation.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—1.2 GB 2.1 GB
Fits Get file →

The available quantization is IQ4_X_S, which is a 4‑bit high‑quality quant. The file size is 1.2 GB and requires about 2.1 GB of VRAM at 8k context. For lower VRAM budgets, consider a higher‑bit quant if available, but no other quants are listed.

What GPU do you need for BlazerApex Vision V4 MM?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is the main file (1.2 GB file, about 2.1 GB in use).

Specifications

Architecture from the GGUF header

Architectureqwen35
Parameters1.88B
Layers24
Trained context262,144 tokens
KV cache (fp16)48 KB per token

Source

Quantized byDavizig10jojo
Licenseapache-2.0
Downloads (HF)116
Last updated2026-10-09

Frequently asked questions about BlazerApex Vision V4 MM

How much VRAM does BlazerApex Vision V4 MM need?

About 2.1 GB at its main file with an 8K context window. That is the 1.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run BlazerApex Vision V4 MM on an 8 GB GPU?

Yes, with the smaller files: the 1.2 GB file needs about 2.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of BlazerApex Vision V4 MM GGUF?

It depends on the quant: file is 1.2 GB.

What context length does BlazerApex Vision V4 MM support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.4 GB at 8K and 1.6 GB at 32K.

Where can I download BlazerApex Vision V4 MM GGUF?

The files are on Hugging Face in the Davizig10jojo/BlazerApex-Vision-V4-MM-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does BlazerApex Vision V4 MM use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run BlazerApex Vision V4 MM on 8 GB VRAM?

Yes, the IQ4_X_S quant uses about 2.1 GB of VRAM at 8k context, leaving room for the mmproj file and other overhead.

Does the model support vision tasks?

Yes, but you must load the mmproj-BlazerApex-Vision-V4-MM-F16.gguf projector file to process images.

Is the model licensed for commercial use?

It is released under the Apache‑2.0 license, which allows commercial use with attribution.

Similar models

Browse all →

More from Davizig10jojo

All by Davizig10jojo →

More Blazerapex models

All Blazerapex →

Other tiny models

All tiny models →