When and how to use Lara Think

Lara offers two translation modes with different levels of reasoning effort:

Lara Base is the default translation mode, balancing quality, speed and cost. It is recommended when low latency is important.
Lara Think applies deeper linguistic reasoning to help improve accuracy and consistency, with longer processing times and a higher cost per source character.

Lara Think is available for API text translation on Pro and Team plans.

What is Lara Think?

Lara Think is Lara's reasoning mode. It applies deeper linguistic analysis before returning the final translation.

It uses multi-step linguistic analysis to refine the translation using the source text and the context you provide. You can use glossaries and translation memories to guide terminology and adaptation.

Lara Think can reduce the amount of post-editing needed. The benefit depends on the content, the context provided and your review requirements.

When should I enable Lara Think

Choose based on the needs of each workflow. Lara Base is also suitable for professional translation; enable Think when the potential benefit of deeper analysis outweighs the extra processing time and cost.

	Lara Base	Lara Think
Focus	Quality, speed and cost	Deeper reasoning for accuracy and consistency
Best for	General-purpose translation and real-time applications	Professional workflows where deeper analysis justifies added time and cost
How it works	Fast, high-quality translation	Multi-step analysis of the source text and available context
Processing time	Fast, suited to real-time use	Can take up to 10 times as long as Lara Base
Glossaries and translation memories	Supported	Supported
Plans	Free, Pro, Team	Pro, Team
Pricing	Standard API rate per source character	Separate API rate per source character
How to use it	Default mode; reasoning is off	Set reasoning to true in your SDK request

Getting the best results from Lara Think

Less revision work. Lara Think reduces the amount of content that needs to be revised. Where human review is still part of your workflow, that review follows human timelines. The time-to-market gain comes from having less to revise, not from instant delivery.

Context still matters. Lara Think makes the most of the information you give it. If a source text is ambiguous and no context is available, no model can reliably choose the right interpretation. In these cases, the most effective way to improve results is to add context, glossaries or instructions.

How to use Lara Think

Enable reasoning for API text translation by setting the reasoning option to true in your SDK request.

reasoning: true (in Python: reasoning=True)

For supported SDKs and complete code examples, see Translate Text in the developer documentation.

Where your SDK supports partial results, you can use a callback to receive a preliminary translation while reasoning continues. The final response contains the refined translation.
More info here.

Pricing

As with all of Lara's text translation services, you are charged only for the characters in the source text. Lara Think requires more computation, so it is billed at a higher rate than Lara Base. For current rates, see the Pricing page in the developer documentation.

Frequently asked questions

Does Lara Think guarantee an error-free translation?
No. Review requirements depend on your content and intended use.

Can I use my glossaries and translation memories with Lara Think?
Yes. You can pass the glossaries and translation memories you use for API text translation when enabling reasoning.

How is Lara Think different from Lara Prosa?
Lara Think adds reasoning to translation. Lara Prosa is a separate service that refines translations against a style guide, with a focus on editorial style and brand voice.

Why does Lara Think take longer?
Deeper reasoning requires additional computation. Use Lara Base when low latency is a priority.

This article is about:

Lara Think
Lara Base vs Lara Think
Reasoning in the Lara API
Professional translation and localization workflows