Best AI models by task

Large reasoning models handle multi-file refactors, debugging and agentic coding best. Compare the strongest coding-capable chat models side by side with live prices.

For autocomplete, small edits and high-volume code tasks, smaller models give most of the value at a fraction of the price.

Multi-step analysis, maths and planning need models that think before answering. These are the deepest-reasoning models available on TokenRex.

Support bots need fast, consistent and affordable replies at high volume. These balanced models are a good starting point.

Contracts, research papers and large PDFs need long context windows. Compare models that read large inputs, with live input prices — the biggest cost driver here.

Autocomplete, live assistants and UI helpers need replies in milliseconds. Fast-priority versions route your request ahead of the queue for about twice the price.

The lowest-cost chat models on TokenRex, sorted by live input price.

Compare text-to-image models from OpenAI and Google with live per-image prices.

Low-cost image models for thumbnails, drafts and bulk generation.

Generate short clips from text or images with Veo and Gemini Omni, billed by length.

Natural voices for narration, apps and accessibility, compared by live price.

Transcribe meetings, calls and voice notes. Compare transcription models by live price per unit.

Embeddings power semantic search and retrieval-augmented generation. Compare OpenAI and Google embedding models by live price.

Build phone-style voice assistants with live audio models billed per minute.

Generate music clips and tracks from text with Lyria, billed per track.

Agents call tools, plan over many steps and recover from errors. Strong tool-using models keep agents on track.