← Catalog
K-EXAONE 2.5
AvailableLG AI ResearchProprietary
LG AI Research's proprietary foundation model, a technical-validation release built by further training the K-EXAONE model released in August. Text-only, always reasons before answering (reasoning_effort low/high/max), 524,288-token context. LG-reported agentic scores: AutomationBench-Openset 74.6, SWE-Bench Pro 59.2, BrowseComp 78.0, Claw-Eval 77.3. API offered free of charge on the EXAONE API Platform from Nov 9, 2026.
Specifications
- License
- Proprietary
- Weights
- Not released
- Architecture
- unknown
- Parameters
- Undisclosed
- Context window
- 524K tokens
- Max output
- —
- Knowledge cutoff
- —
- Price (in / out, $/M)
- —
- Modalities
- Text
Benchmarks
No benchmark scores recorded yet. Spotted some? Submit a correction.
Vendor-reported figures are claims until independently verified. See methodology.