LLM models

A powerful AI reasoning model that outperforms OpenAI O1 in complex tasks. This lightweight version excels in mathematics, analysis and problem solving with remarkable efficiency

This new AI model from OpenAI has been reinforcement-trained for complex reasoning. It can think internally before answering you. It outperforms humans in some difficult tests

Solve complex theorems in Lean 4 with an open-source model combining informal reasoning and formal proofs. Break down problems into sub-objectives and generate detailed solutions.

A family of fully open language models with source code, training data and transparent evaluations. Ideal for researchers and developers wishing to explore and improve AI

A diffusion-based language model that generates text up to 10 times faster than traditional LLMs. Benefit from increased efficiency and reduced cost for your AI applications requiring fast responses

A 72-billion-parameter multimodal model that excels in visual reasoning. Solve complex math and physics problems with a deep understanding of images

A multimodal model that makes images and text much easier to understand. Analyze your visual documents, answer complicated questions and generate detailed descriptions via 3 MLLM models

Enjoy excellent value for money with 2M contextual tokens, integrated web/X search, and fast reasoning. Remarkable performance at low cost for development and analysis

Enjoy even more human-like and intelligent conversational AI: GPT-5.1 responds faster, follows your instructions better, adapts its tone to your preferences, and handles complex or creative tasks with natural warmth.…

OpenAI's low-cost but powerful LLM model: multimodal reasoning, computer coding, long-form writing, etc. Ideal for chatbots, data analysis and task automation

An LLM model that looks very promising ( not much information yet). You can test it directly on the Chatbot Arena platform.

A multimodal model that rivals GPT-4 and Claude 3 in terms of performance. It includes images, video and sound. It can reason, code and speak 32 languages