Best open-weight chat model
European AI lab producing frontier open-weight models deployable on your own infrastructure with no usage fees and no data leaving your environment. Fast, efficient, and competitive with much larger proprietary models on most benchmarks. The default choice for teams that want serious AI capability without vendor lock-in or compliance headaches.
Pricing plans
| Directr tier | B | S | S | S |
| Price | Free / API usage-based | Free / $20/mo Plus | Free / $20/mo Pro | Free / API from $0.028/1M tokens |
| Free tier | Free | Free | Free | Free |
| API access | ✓ | ✓ | ✓ | ✓ |
The tool that made AI assistants mainstream — and still the most broadly capable for everyday use. Best-in-class for general knowledge, coding, content drafting, data analysis, and image generation via DALL·E 3. Four years of model improvement and a vast plugin and GPT ecosystem give it a feature lead that's hard to catch up to.
Anthropic's AI assistant — the one most often cited by professionals who care about depth. Exceptional at long-document analysis, nuanced conversation, research synthesis, and tasks requiring genuine reasoning rather than pattern-matched confidence. Widely used by writers, analysts, lawyers, and engineers who need an assistant that's careful, not just fast.
A Chinese-developed frontier model matching GPT-4o quality on most benchmarks — and the R1 reasoning model is a genuine rival to OpenAI's o1 for mathematical and logical problem-solving. Free to use at deepseek.com, with an API priced at roughly one-tenth of OpenAI's rates and full OpenAI compatibility for painless migration.