Models in Prompt Compare

Why these models?

Prompt Compare focuses on widely used AI platforms people actually evaluate for work — not every model on the internet. Each selectable model is wired through the Vercel AI Gateway and must stay within our per-session compute guard (roughly at or below Anthropic Fable 5 lane pricing at max output).

Each session uses official model APIs so you see real output, not screenshots from a chat UI.

Why some platforms are not listed

• No public chat window — cloud ML consoles (AWS SageMaker, Azure ML, Bedrock playground-only) are operator tools, not consumer chat products.
• API-only or developer gateways — Fireworks, Cerebras, and similar hosts are inference APIs without a ChatGPT-style prompt page for visitors.
• Cost exceeds session guard — frontier models above our Fable 5 lane budget appear grayed out in the tool (for example GPT 5.4 Pro).
• Not on the gateway roster — if Vercel AI Gateway does not expose a stable model ID, we cannot run it in a $6 live comparison yet.
• Aggregator wrappers — Poe and HuggingChat are useful links on the hub directory, but Prompt Compare tests first-party API models only.

Want another model in the tool? Suggest an AI for our tool — same inbox as the hub companies page (feedback@ainerdnetwork.com).