HyperCLOVAX-SEED-Think-14B-GPTQ
by Naver
HyperCLOVAX-SEED-Think-14B-GPTQ is a 4-bit quantized version of Naver’s HyperCLOVAX-SEED-Think-14B, a 14B-parameter causal language model optimized for Korean and English conversational tasks. Quantized using GPTQ (W4A16g128), it retains strong performance (e.g., MMLU: 0.7018, HAERAE: 0.8139) while reducing computational and memory requirements. Developed by K-Compression (quantization) and Naver HyperCLOVA (base model), it is licensed under Naver’s proprietary HyperCLOVA X SEED Model License Agreement and is ideal for cost-effective deployment in multilingual applications. Requires GPU support for inference.
Transparency Score
Weighted across four pillars · updated May 4, 2026
Vendor Information
Complete information about the vendor/provider of this AI application
EU AI Act Provider Information
Get insights into risk by running assessments on this AI application.
Work at Naver? Claim this listing to correct or complete the data.
Ready to manage AI applications?
Track, assess, and govern your AI applications with Anove.