Search the full model catalog.

The 0.5B-parameter chat model of Qwen1.5, the smallest size in the series, post-trained with supervised fine-tuning and DPO. It suits short replies and rewrites where the prompt supplies the facts.

Microsoft's 14B-parameter open model for English math and reasoning, built for memory- or compute-limited setups and trained with synthetic textbook-style data. MIT licensed.

The 32.5B-parameter flagship of Qwen2.5-Coder, instruction-tuned for writing and repairing code. Qwen reports 73.7 on the Aider code-repair benchmark and 65.9 on McEval, which spans more than 40 languages.

Gemma 2 2B IT is the smallest instruction-tuned Gemma 2 model from Google, with about 2.6B parameters, for short English replies and rewrites where the prompt supplies the facts.

Gemma 2 9B IT is Google's 9-billion-parameter instruction-tuned Gemma 2 model for English chat and writing. Google reports its pretrained base at 68.6 on GSM8K math, against 23.9 for the 2B.

Meta's 8B-parameter Llama 3 model tuned for English assistant chat. It answers with text and code inside an 8k-token context and can keep a separate chat history per user and session.

Mistral AI's 7B chat model, fine-tuned for instructions from Mistral-7B-v0.3. This version added function calling and moved to the v3 tokenizer with a 32,768-token vocabulary.

The 8B instruction-tuned model of Meta's Llama 3.1 collection, for chat and code in eight languages. It was pretrained on roughly 15 trillion tokens, and its knowledge stops at December 2023.

Mid-size Qwen2.5 chat model (7.61B parameters) with JSON mode, function calling and up to 32,768 input tokens on Wiro. Qwen reports 75.5 on MATH and 84.8 on HumanEval.

Wiro AI's 8B Turkish chat model: Llama 3.1 8B Instruct fine-tuned on more than 500,000 Turkish instructions, with a focus on Turkish idioms and local context.

Detects the faces in a photo and hides each one under a heavy oval blur, leaving the rest of the image untouched. One image in, one PNG out at the original size.

The pretrained 13B-parameter Llama 2 base model from Meta, with no chat or instruction tuning. It continues the text you give it, for story drafting and pattern completion in English.

This is FLUX.1-dev-fp8 v1.0 model.

Answers questions about one image or captions it, using Google's instruction-tuned Flan-T5-XL as the language model inside BLIP-2. Only the answer text comes back.

The larger BLIP captioning checkpoint from Salesforce, with a ViT-L image encoder and trained on COCO. It takes one image plus an optional opening phrase and returns a short English caption.

BLIP captioning checkpoint from Salesforce with a ViT-B image encoder, trained on COCO. It writes a short English caption for an image and can continue an opening phrase you give it.

This is an FLUX1.0 image to image tool.

Inpainting and outpainting model from Black Forest Labs. Upload a photo and a mask, describe what belongs in the masked area, and only that area is repainted to fit the rest of the image.

LoRA training is a low-rank adaptation technique used to fine-tune FLUX models, enabling personalized model outputs with minimal computational overhead.

Writes a detailed English description of a video clip, guided by your prompt. Built on THUDM's CogVLM2-Llama3-Caption, the model that captioned training data for CogVideoX.