Skip to main content
OpenAI provides the widest model ladder in the translation catalog: seven selectable tiers. Within this family the tier choice matters far more than the family choice. The current generation is GPT-5.6, which ships as Luna, Terra, and Sol in ascending capability. All three carry the same 1.05M-token context window and 128K maximum output — the tiers differ in capability, not in how much text they can hold. Ollang exposes the two ends of that range.
Available in Ollang Workflows — all OpenAI tiers serve Document Translation, Subtitle Translation, and AI Dubbing. See the Translation catalog.

​
Models

​
Choosing a tier

GPT-5.6-Sol is OpenAI’s most capable model, for complex translation and localization where nuance, terminology accuracy, consistency, and reliable instruction-following all have to hold simultaneously. It supersedes both GPT-5.2 and GPT-5.4 on capability, which makes the older 5.x tiers hard to justify for new Workflows. GPT-5.6-Luna is the efficiency tier of the same generation, while carrying the identical 1.05M-token context window. That combination is unusual: whole-document context on a throughput-oriented tier. It is the right default for high-volume, quick-turnaround content. GPT-5.4 brought a roughly 1M-token context window — more than double GPT-5.2’s 400K — along with configurable reasoning effort. Worth keeping selected only where an existing benchmark depends on it. GPT-5.2, GPT-5, GPT-4.1, and o4-mini are earlier tiers. GPT-4.1 remains a sensible mid-tier workhorse with 1M-token context; o4-mini is the compact reasoning option, with an efficient tokenizer that keeps non-English text economical to process.

​
Where it fits in a workflow

A common two-tier configuration pairs GPT-5.6-Luna as the Global Workflow default with GPT-5.6-Sol overridden at Folder level for customer-facing or regulated content. Because both tiers share a context window, the throughput tier gives up nothing on document length.

​
Trade-offs

  • The tiers between GPT-5 and GPT-5.4 are now largely superseded by GPT-5.6-Sol.
  • Model names and defaults change frequently; confirm current behavior against OpenAI’s model documentation before standardizing a Workflow on a specific tier.

​
Reference