GPT-4o ("o" for omni) is OpenAI's multimodal flagship of the 4-series and the default pick for the OpenAI provider on Magnus. It is a polished all-rounder with particularly good image understanding.
Launched May 13, 2024 by OpenAI as the first natively multimodal GPT-4-class model - one network trained across text, vision, and audio rather than separate systems stitched together.
On the All-inclusive plan a GPT-4o reply costs 3 credits.
General chat and anything involving images - drop a screenshot, a whiteboard photo, or a chart into the conversation and GPT-4o reads it well. With an OpenAI key on file it also unlocks the OpenAI speech engine in voice settings.