Modèles
Language models you can run or call, from open-weight releases to hosted frontier systems. Compare them by context window, licence, parameter count and cost.
16 technologies
- LLM concept Large language model: a Transformer trained on text at a scale that yields general capability.
- GPT model Family of decoder-only Transformer language models trained with a generative objective.
- Llama model Open-weight language model family from Meta, released for local and self-hosted use.
- BERT model Encoder-only Transformer pre-trained with masked language modelling.
- Qwen model Alibaba's open-weight model family, strong on multilingual and coding tasks.
- Mistral model European open-weight model family known for efficiency at small parameter counts.
- Claude model Anthropic's hosted model family, positioned around long-context work and safety.
- DeepSeek model Open-weight model family notable for efficient training at frontier scale.
- Hallucination concept Fluent output that is not supported by the model's input or by fact.
- Gemma model Google's open-weight model family, released in small sizes.
- Gemini model Google's hosted multimodal model family, integrated with its cloud platform.
- Token concept The unit of text a language model reads and writes.
- Logits concept The raw, unnormalised scores a model produces before softmax.
- Phi model Microsoft's small models trained mainly on curated textbook-quality data.
- Context Window concept Maximum number of tokens a model can attend to in one request.
- Command R model Model family oriented toward retrieval and tool use.