Mistral: Pixtral 12B

Mistral Small 3.1 24B, released on March 17, 2025, is an open-weight multimodal model from Mistral AI, distributed under the Apache-2.0 license. With around 24B parameters and a 128K token context window, it is available in both base and instruction-tuned (“Instruct”) variants. The model introduces vision support alongside text, enabling tasks like multimodal reasoning, captioning, and image-based Q&A.It is multilingual, supporting many languages, and is optimized for fast responses, function calling, structured dialogue, and long-context reasoning. Despite its size, the model can be run locally in quantized formats, fitting on machines with ~32GB RAM, making it accessible to developers outside large cloud setups. However, the output length is smaller than the 128K input window, meaning long generations may require chaining. In addition, using full vision features or the maximum context window significantly increases compute costs, and performance on highly complex reasoning or enterprise-scale tasks still trails larger proprietary frontier models.

Mistral Medium 3.1, released in August 2025 as the mistral-medium-2508 update, is a proprietary frontier model from Mistral AI positioned between smaller open models and high-end closed LLMs. It is multimodal, handling both text and image inputs, with a context window of ~128K tokens. Compared to Mistral Medium 3.0, the 3.1 release introduces improvements in reasoning, coding, STEM, and enterprise workflows, along with better tone control for conversational and business applications. It is designed for scalable enterprise deployments, including hybrid cloud and on-premises VPC setups. As part of Mistral’s Premier line, Medium 3.1 is a commercial-only offering: while it delivers strong accuracy and performance, trade-offs include higher costs than open-weight models, restricted fine-tuning access, and increased latency/cost for very large contexts.

Qwen3 VL 30B A3B Instruct is an open-weight multimodal large language model developed by Alibaba as part of the Qwen family, built for instruction-following tasks that unify text generation with visual and video understanding. Released around October 2025 under the Apache-2.0 license, it targets efficient, high-fidelity vision-language reasoning across very long contexts.The model accepts text and image inputs and produces text outputs, with strong performance in OCR, spatial reasoning, long-video understanding, and agentic or GUI-centric visual tasks. It uses a Mixture-of-Experts (A3B) design with ~31.1B total parameters and ~3B active per token, paired with Qwen3-VL’s unified multimodal stack (including Interleaved-MRoPE and DeepStack fusion) to process text, images, and video in a single architecture. OCR support expands to 32 languages, enhancing document workflows. With a native ~262K token context window (extendable further), it stands out today for its balance of scale, efficiency, long-context support, and open accessibility in multimodal systems.

Gemma 3 12B, announced by Google DeepMind on March 12, 2025, is part of the open-weight Gemma 3 family, designed to provide a balance between capability and accessibility. With around 12 billion parameters, it supports multimodal input (text + images) and outputs text, making it useful for reasoning, summarization, Q&A, and visual understanding tasks. The model supports an input context of 128,000 tokens and typically generates up to ~8,000 tokens in output.The 12B variant is instruction-tuned (“Gemma-3-12B-IT”) and optimized for multilingual use across more than 140 languages. It can run on a single GPU or TPU, offering a lighter compute footprint than very large proprietary models, while still achieving strong performance in reasoning benchmarks. Quantized and lower-precision variants are available to improve efficiency. Limitations include smaller output lengths relative to input capacity, scaling hardware needs at larger sizes, and performance below massive proprietary models on the most complex multimodal or reasoning-heavy tasks.

Gemma 4 12B is an open-weight multimodal model from Google in the Gemma 4 family. It is intended for text and image understanding tasks such as visual question answering, OCR, captioning, and document understanding, with a smaller parameter footprint than the larger Gemma 4 variants.This entry is connected to Roboflow Playground vision evals for comparison. No runnable Playground workflow is configured yet, so the model page is used for discovery and benchmark context rather than direct hosted inference.

Llama 4 Scout, released on April 5, 2025, is one of Meta AI’s first Llama 4 multimodal models, alongside Maverick. It accepts text + image inputs and produces text outputs, with a knowledge cutoff of August 2024. Scout is notable for its extremely large context window of 10 million tokens, making it well-suited for analyzing very long documents, extended conversations, or large codebases.Architecturally, Scout uses a Mixture-of-Experts (MoE) system with 16 experts, activating ~17B parameters per inference from a pool of ~109B total parameters, balancing capacity with efficiency. It officially supports 12 languages (including English, Arabic, French, Hindi, and Spanish), while offering multimodal reasoning for images (captioning, Q&A, recognition). Meta highlights that Scout can run on a single Nvidia H100 GPU, making it more accessible than larger-scale Llama 4 models. However, its output token limit is far smaller than its 10M input window, image input support is still constrained, and license restrictions apply for large-scale commercial deployments.