Open-source AI models,
compared at a glance.

Distilled BERT model that encodes sentences into 384-dimensional vectors for measuring semantic similarity. Trained on over a billion sentence pairs spanning scientific papers, web QA, NLI datasets, and community forums. At 22M parameters and 6 transformer layers, it is fast enough for CPU inference while remaining competitive on standard sentence similarity benchmarks.

Qwen3-VL-2B-Instruct is a 2-billion-parameter vision-language model from Alibaba Cloud that jointly processes images and text for visual question answering, captioning, and document understanding. Its 2B scale positions it as one of the smaller instruction-tuned VLMs capable of zero-shot visual reasoning. Apache 2.0 licensed.

Google's original BERT base model in uncased form, pre-trained on BookCorpus and English Wikipedia via masked language modeling. Tokens are lowercased before processing, making it insensitive to capitalization. It remains a standard fine-tuning base for classification, NER, and extractive QA, though newer encoders outperform it on most benchmarks.

ELECTRA base discriminator from Google, pre-trained using replaced token detection rather than masked language modeling. A small generator produces candidate replacements; this model learns to identify which tokens were swapped — a task that uses every token for training signal, making pre-training more efficient than BERT per compute dollar. Intended as a fine-tuning base for classification and token-level tasks.

Multilingual sentence embedding model covering 50+ languages, built on a 12-layer distilled MiniLM architecture. Produces 384-dimensional vectors designed for semantic similarity and paraphrase detection across language boundaries. Trained on multilingual paraphrase data to align semantically equivalent sentences even when expressed in different languages.

Cross-encoder reranker trained on the MS MARCO passage retrieval dataset, designed to score query-document pairs jointly rather than encoding them independently. Distilled from a 12-layer cross-encoder into 6 layers to reduce latency while retaining re-ranking accuracy. Used as a second-stage ranker on top of fast first-stage retrieval (BM25 or bi-encoder).

Sentence embedding model based on the MPNet architecture, producing 768-dimensional vectors. Trained on over a billion sentence pairs from MS MARCO, NLI datasets, and community QA forums, it is frequently used when accuracy matters more than inference speed among English embedding models. The MPNet backbone enables masked and permuted prediction during pre-training for stronger representations.

Small English dense embedding model from BAAI's BGE (BAAI General Embedding) series, producing 384-dimensional vectors via MIT license. Optimized for MTEB retrieval benchmarks through a retrieval-focused training strategy, it achieves competitive scores relative to its parameter count. Suited for embedding workflows where throughput and cost matter more than peak accuracy.

OpenAI's CLIP model using a ViT-L/14 image encoder, trained contrastively on 400 million image-text pairs from the internet. It aligns image and text in a shared embedding space, enabling zero-shot image classification by comparing image embeddings against text label embeddings. The ViT-L/14 variant offers higher accuracy than the smaller ViT-B/32 at greater compute cost.

MobileNetV3 small model at 100% width multiplier, trained on ImageNet-1k using the LAMB optimizer via the timm library. At under 3M parameters, it targets image classification on mobile and edge hardware where latency and memory are primary constraints. Part of timm's standardized pretrained model zoo with consistent preprocessing and inference APIs.

Chronos-2 is Amazon's second-generation pretrained foundation model for zero-shot time-series forecasting. It frames forecasting as a language modeling problem over quantized time-series tokens using a T5 encoder-decoder architecture, enabling it to forecast across diverse domains without per-dataset training. Released under Apache 2.0.

Vision Transformer (ViT) fine-tuned for binary NSFW vs. safe image classification. Provides a single classifier for flagging potentially unsafe image content without category-level labeling. Built on ViT-base architecture and fine-tuned on a curated dataset of safe and unsafe images.

OpenAI's CLIP model using a ViT-B/32 image encoder, the smaller of the two widely deployed CLIP variants. Trained contrastively on 400 million image-text pairs, it aligns image and text representations in a shared embedding space for zero-shot classification and retrieval. The B/32 variant sacrifices accuracy versus ViT-L/14 for faster inference.

BAAI's BGE-M3 embedding model supporting over 100 languages with a unified architecture capable of dense, sparse (lexical), and late-interaction (ColBERT-style) retrieval modes from a single checkpoint. Built on XLM-RoBERTa with large-scale multilingual training, it targets multi-lingual and cross-lingual retrieval where a single model must handle diverse language inputs.

Qwen3-0.6B is the 0.6-billion-parameter instruction-tuned model from Alibaba Cloud's Qwen3 series, fine-tuned from the Qwen3-0.6B-Base for conversational and task-following use. It targets deployment in environments where even a 1B model is too large — edge hardware, mobile devices, or ultra-low-latency services. Apache 2.0 licensed.

RoBERTa base from Facebook AI, trained with the same architecture as BERT base but significantly more data, longer training schedules, larger batch sizes, and dynamic masking. Pre-trained on BookCorpus, Wikipedia, CC-News, OpenWebText, and Stories — substantially more data than the original BERT. MIT licensed with multi-framework support.

RoBERTa large, the 355M-parameter version of Facebook AI's strongly trained BERT variant, offering doubled hidden size and additional attention heads over RoBERTa base. It provides stronger NLU accuracy at roughly 4x the inference compute cost of the base variant. Used where task accuracy on complex English language understanding outweighs latency constraints.

XLM-RoBERTa base from Facebook AI, pre-trained on 2.5TB of filtered CommonCrawl text across 100 languages using the RoBERTa training procedure. Enables cross-lingual transfer — models fine-tuned on labeled English data can infer on other languages without parallel annotations. The standard starting point for multilingual classification and token-level tasks.

LAION's CLAP (Contrastive Language-Audio Pretraining) model using the HTSAT (Hierarchical Token-Semantic Audio Transformer) encoder, fused with a text encoder to align audio and text in a shared embedding space. Analogous to CLIP for images, it enables zero-shot audio classification and retrieval using natural language descriptions without task-specific labeled audio data.

OpenAI's original GPT-2 at 124M parameters, an autoregressive language model trained on WebText (over 8 million web documents filtered from Reddit outlinks). It generates English text continuation given a prompt using next-token prediction, trained without any instruction tuning or RLHF. MIT licensed and runnable on commodity CPU hardware.

ADetailer is a collection of Ultralytics YOLO-based face, body, and hand detection models distributed for use with the Stable Diffusion WebUI's ADetailer extension. The models detect regions of interest in generated images (faces, hands) to trigger targeted inpainting passes for quality improvement. Trained on WIDER FACE and anime segmentation datasets, covering both photorealistic and anime styles.

Nomic Embed Text v1.5 is a matryoshka-capable English embedding model from Nomic AI, built on a custom nomic-BERT architecture trained with contrastive learning on large-scale text pairs. Matryoshka Representation Learning allows truncating embeddings to shorter dimensions (e.g. 64, 128, 256) without retraining, enabling flexible precision-cost tradeoffs. The model is transformers.js-compatible for browser-side inference.

BGE-Large-EN-v1.5 is BAAI's highest-capacity English embedding model in the v1.5 series, producing 1024-dimensional vectors. It achieves top MTEB retrieval scores among its generation of English-only embedding models, at the cost of higher compute and storage than BGE-small or BGE-base. MIT licensed with ONNX export support.

ColBERTv2 is a late-interaction retrieval model from Stanford that encodes queries and passages as per-token embeddings rather than a single vector, allowing MaxSim matching at retrieval time. This token-level interaction yields higher accuracy than bi-encoders on many retrieval benchmarks while remaining more efficient than cross-encoders. The model is MIT licensed and implemented in PyTorch with ONNX support.