Skip to content
← Models
Open weightsKorea University NLP & AI Lab/10 trustMetrics-checked

KULLM

Korean instruction-tuned LLM family from Korea University researchers.

Visit official page →

About

KULLM is a Korean large language model family developed by Korea University's NLP and AI Lab with the HIAI Research Center, first introduced in June 2023. The project built high-quality Korean instruction datasets and fine-tuned models including Polyglot-ko variants and later KULLM3 on Upstage's SOLAR-10.7B, targeting practical Korean language generation for chatbots and RAG question answering.

The team reported Korean conversational performance competitive with GPT-3.5 Turbo on fluency, coherence, and accuracy using GPT-4-based evaluation, though these are team-reported numbers. Checkpoints are public on Hugging Face. License terms for specific variants should be verified before commercial use. Release date is era-level (June 2023 debut).

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

June 1, 2023

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

korean-nlpchatbotsinstruction-followingresearchrag

Trust score — 6/10

/10 trustMetrics-checked

Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 119; Trending score 62/100 from researcher assessment.

Sources