← Back to docs
Which model should I use?
Match your task to the right model. Start with the use case, then pick the recommended option.
Chat
gpt-4oFast, multimodal, great for conversation and general tasks.
Reasoning
gpt-5-codexDeep logic, math, and multi-step problem solving.
Cheap / high-volume
Llama-4-MaverickOpen-weight, cost-effective for bulk inference.
Vision
gpt-4o or claude-opus-4-8Image understanding, OCR, and visual Q&A.
Speech
whisper-1 / tts-1Transcription and text-to-speech pipelines.
Still unsure? Start with gpt-4o — it handles most tasks well. Switch to a specialized model when you hit a ceiling.