Skip to content
Computer vision

Inkling 2 Small

open-weights multimodal model

2 min readupdated

Inkling 2 Small is a general-purpose multimodal model that takes text and visual inputs and generates text output. It is built for developers creating AI-powered applications, including agentic systems, chatbots, and coding tools. The model uses a 48-layer decoder with a sparse Mixture-of-Experts (MoE — an architecture where only a subset of parameters is activated per token) and supports a context window of up to 1 million tokens.

Key features

  • Supports text and image inputs processed together in a shared latent space.
  • 594 billion total parameters, with 17 billion active per token.
  • Hybrid attention combining local and global layers.
  • Native support for NVFP4 and MXFP8 formats.

What you can do

  • Build smart chatbots and agents for complex multi-step tasks using images.
  • Integrate the model into IDEs for code analysis and autocompletion.
  • Set up retrieval-augmented generation (RAG) systems to work with large documents.

Example

import torch
from transformers import AutoModelForVision2Seq, AutoProcessor

model_id = "crusoeai/Inkling-2-Small-MM"
processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForVision2Seq.from_pretrained(model_id, torch_dtype=torch.bfloat16, device_map="auto")
Read moreDetails
Benchmarks: AA Index — 30.6, Vals Index — 56.1%, FrontierCode v2.0 — 25.4%
Hardware requirements: at least 500 GB VRAM for NVFP4 (for example, 2x NVIDIA B300 or 4x NVIDIA H200)
Highlights: natively trained in NVFP4, supports integration via vLLM, SGLang, and Unsloth
Discuss Telegram 0 21

Will it run?

Won't run

Not enough memory even for 4-bit. Below are GPUs you can rent.

  • U8weights 560 GB · ready files616 GB
  • FP8 / INT8recommendedweights 553 GB609 GB
  • 4-bitweights 311 GB343 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

No single instance holds the whole model: it needs RAM offload or a cluster.

Your own GPU: Won't fit any home GPU — renting is cheaper.

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # crusoeai/Inkling-2-Small-MM
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download crusoeai/Inkling-2-Small-MM --local-dir ~/models/Inkling-2-Small-MM
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to Inkling-2-Small-MM
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of Inkling-2-Small-MM (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

crusoeai/Inkling-2-Small-MM
Author
crusoeai
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

210

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments