exaone-deep:7.8b
Upstream reference: EXAONE Deep 7.8B
Publisher: LG AI Research
LG AI Research documents a reasoning-oriented derivative of EXAONE 3.5 for English and Korean mathematical and coding text.
Publisher-described context, separate from Sigma5C observations and recorded catalog statistics.
Published upstream specifications
- Parameters excluding embeddings
- 6.98 billion Source
- Published context
- 32,768 tokens Source
- Upstream license metadata
- EXAONE custom license; review linked terms Source
Publisher-described uses
- Reasoning-oriented text research and the documented mathematical/coding prompts. Source
Reference limitations
- The parameter count excludes embeddings; the card’s custom terms are not a generic permissive-license claim. Source
A recorded tag and an upstream card association do not verify the deployed checkpoint, quantization, configured context or current hosted availability.
Publisher sources
- EXAONE Deep 7.8B official documentation — LG AI Research. Checked .
Links
Recorded references do not establish downloadable weights, licensing or current availability.
7.8B
parameters
Q4_K_M
quantization
exaone
architecture
32768
context
EXAONE AI Model License Agreement 1.1 - NC
license
2025-03-18
released
Where exaone-deep:7.8b comes from
Company / lab
LG AI Research
See every model and family from this maker.
Model family
exaone-deep
Other versions and sizes in this line.
completionthinking