Every model, one platform.

Cicle's own model, hosted in Germany. In development — not yet available to select.

OpenAI's newest release of GPT-6 Sol, at the same price as Sol itself.

OpenAI's balanced GPT-6.1 tier: strong reasoning and clear writing at half the price of 6.1 Sol.

OpenAI's frontier generation at its most capable, for project work with long documents and careful reasoning.

OpenAI's GPT-6 generation at a low price, for everyday chat and high-volume routine work.

OpenAI's deepest reasoning tier, for analysis worth waiting a little longer for.

The balanced frontier model: strong reasoning without the top-tier price.

A very large context window at a mid-range price, for threads and documents that will not fit elsewhere.

The previous frontier generation, retained for setups already tuned to it.

Extended reasoning from the 5.5 generation, for long analytical chains.

An earlier balanced model, retained so existing configurations keep resolving.

Fast to the first word and inexpensive — the everyday choice for chat.

The lightest 5.4 tier, retained for simple classification at scale.

The cheapest capable tier, for high volume where cost matters more than speed.

The smallest, cheapest tier, for classification and short mechanical tasks.

Mistral's flagship: European-built, and strong on multilingual work.

The balanced Mistral — much of Large's quality at a lower price.

Mistral's compact tier, for fast, inexpensive, high-volume turns.

Anthropic's long-form writing model, tuned for narrative and tone.

Anthropic's most capable model, for complex engineering and agentic work.

The balanced Claude: close to Opus quality at a working-day price.

The previous Opus generation, retained for configurations pinned to it.

A dependable all-rounder for drafting, summarizing, and tool use.

The fastest and cheapest Claude, for short turns at high volume.

Google's fast tier: inexpensive, current, and comfortable with long documents.

The previous Flash generation, retained for setups already using it.

Google's top reasoning tier. Still in preview — expect it to change.

The lightest Gemini, for bulk classification and extraction.

An earlier Pro generation, retained so existing setups keep resolving.

xAI's current frontier model, its most capable general-purpose tier.

The balanced Grok, for general assistant work at moderate cost.

An early xAI model aimed at software and build workflows.