Start in the model directory, then open a model page to compare its pricing, providers, benchmark results, latency signals, and compatibility details. The information shown varies with the data available for that model.
Practical answers for comparing models, understanding the data, and using the gateway. Cannot find an answer? Reach out to our team.
Start in the model directory, then open a model page to compare its pricing, providers, benchmark results, latency signals, and compatibility details. The information shown varies with the data available for that model.
A model page brings together the useful facts for a specific model: who created it, which providers offer it, the listed prices, benchmark coverage, and gateway availability. Use the page tabs to go deeper into pricing, providers, performance, and quickstart details.
Phaseo stores each provider route's meter, unit size, currency, plan, and effective dates. Input and output rates remain separate, and token prices are normalised per 1M tokens for comparison; they are not combined into a blended headline price. Read the pricing methodology or enter your own token volumes in the pricing calculator.
Gateway availability describes whether Phaseo currently has an eligible route for a model. It is separate from a model's quality or benchmark performance. Provider status can change, so check the model's providers and availability information before relying on a route in production.
Yes. Phaseo Gateway provides one OpenAI-compatible interface for supported models and providers. Create an API key, choose a model, and use the same integration surface while retaining model and provider choice.
Yes. Bring your own provider keys when you want to retain the provider relationship and billing directly, while still using Phaseo's routing, health, and policy layer.
Browse the provider directory for provider coverage, or filter the model directory by capabilities such as chat, embeddings, image, audio, video, and moderation. A provider page is the best place to check its currently listed model catalogue.
Benchmark scores are shown alongside the benchmark and its methodology so that comparisons retain their context. Read how Phaseo normalises AI benchmarks and how it measures latency and throughput before making a production decision from a single score.
Use the contribution page to report a correction or add missing information. Accurate, source-backed reports help keep the directory useful as providers change their catalogues and pricing.