Models & assistants Model or AI service
LG opens its K-EXAONE 2.5 reasoning API
LG AI Research made its proprietary K-EXAONE 2.5 model available through a free, key-based API. It presents this version as a technical validation model. [1]
- Model access
- LG announced free access through its EXAONE API platform with an API key. [1]
- License
- LG describes K-EXAONE 2.5 as proprietary and offers hosted API access. [1]
What the model does
On November 9, 2026, LG announced free API access. Its documentation describes a context limit of 524,288 tokens and lets developers set how much effort the model spends reasoning. [1][3]
Why it matters
A free API lets a team measure how a hosted model handles its own documents and coding tasks before deciding whether to build it into a product. [1][3]
Who can use it
Developers and research teams can obtain an API key from LG’s EXAONE platform. The interface determines how applications send requests and how responses are returned. [1][3]
Key dates
- Free API access announced [1]
Ways to use it
A developer can compare how the model answers a routine request at different reasoning settings. LG documents the response structure and the context limit for requests. [3]
Model limitations
The model always reasons before returning an answer. LG documents a way to change its effort, and the context limit applies to tokens sent in and allowed for the response. [3]
Sources and verification
Catalog evidence checked How we verify records
- Primary source K-EXAONE 2.5 API Available for Free - LG AI Research News
- Independent source LG AI Research opens K-EXAONE 2.5 API free, sees 70% performance gain - CHOSUNBIZ
- Independent source K-EXAONE 2.5 API | EXAONE API Platform
Related profiles
Models & assistants
Hy4
Tencent (Hunyuan)
It is a language model family with documented preview and subsequent variants.
Models & assistants
Qwen4-Max
Alibaba (Qwen team, Alibaba Cloud)
It provides model-based assistance for text and related work tasks.
Models & assistants
Gemini 4 Hydrogen
Google DeepMind
It processes text and other supported inputs for tasks such as code assistance and planning.