Solutions
Pick the best model per task.
Every workload has a different winner. These pages rank the models we route to for each use case, with the auto: directive you can paste straight into your SDK.
Solutions · Coding
Best LLMs for coding
Repo-scale edits, diffs, and green tests on the first try.
auto:coding10 models →
Solutions · Reasoning
Best LLMs for deep reasoning
Frontier math, planning, and multi-step problem solving.
auto:reasoning10 models →
Solutions · Agents
Best LLMs for agents & tool calling
Reliable function calls, JSON mode, and multi-step tool loops.
auto:agent10 models →
Solutions · Long context
Best LLMs for long context
200K to 2M token windows for repos, filings, and transcripts.
auto:long-context10 models →
Solutions · Vision
Best LLMs for vision & multimodal
Charts, PDFs, screenshots, and OCR-grade document parsing.
auto:vision10 models →