kakao/kanana/kanana-2-3b-instruct
Source details
Source revision: d512426765031616ca89fde3a9207f4800d24ccaa319a5dde8d61db0bd4c52b2
Kakao · Text
Kanana models: Versions, specs, GPU requirements & hosting
Kakao Kanana separates instruct and thinking siblings within the same sparse generation for Korean and multilingual reasoning workflows.
Overview and history
Kanana is an official Kakao model family represented here through independently verified release, checkpoint, license, runtime and hardware evidence. Kakao Kanana separates instruct and thinking siblings within the same sparse generation for Korean and multilingual reasoning workflows. Current public releases are grouped by the General, Reasoning tracks so sibling sizes and specialist branches are not presented as false chronological generations.
FusionPointAI currently tracks 4 certified public Kanana releases. The catalogue preserves release dates and evaluates Current, Recent, Previous and Legacy status within each model track rather than by database insertion order.
Instruct and thinking siblings
Sparse deployment profile
Multilingual reasoning
Architecture and model range
The verified public range spans 1.3B to 30B total parameters and context windows up to 32,768 tokens. Each release states whether it is dense or mixture-of-experts and records active parameters separately when the official source publishes them.
Public releases4
Model tracks2
Current generationKanana 2 3B Instruct
Last verified2026-08-26T00:00:00.000000Z
Generations, tracks and variants
General
Kanana 2 2601
Recent generationKanana 2 SLM
Current generation 2 sibling variantsReasoning
Kanana 2 2601
Current generationDeployment ecosystem
Choose a specific Kanana release and certified checkpoint before sizing hardware. FusionPointAI rejects unknown runtime configurations, calculates resident memory from total model weights, then distinguishes globally compatible GPUs from currently rentable inventory.
Validate the exact checkpoint license, runtime version, precision kernel, context length, concurrency and multi-GPU topology. Active MoE parameters affect compute, but resident model memory continues to use total stored weights.
Common use cases
Reasoning assistants
Korean language workloads
Agent evaluation
License overview
Licenses are certified per release, not inferred from the provider. Review the release page before commercial deployment because terms can differ between generations and variants.
Commercial-use status is a source-backed catalogue classification, not legal advice. Follow the linked official license and attribution requirements.
Release timeline
kakao/kanana/kanana-2-1-3b-instruct
Source details
Source revision: 3586f2c262b4dcaac3c2e9dff0fcde120954bbbc03d6491c63749446c6527aa8
kakao/kanana/kanana-2-30b-a3b-instruct-2601
Source details
Source revision: 3499852c92f8b5ca7ce8c17862ff408ecac2013a150e086c7d115f439f6165ca
kakao/kanana/kanana-2-30b-a3b-thinking-2601
Source details
Source revision: d346644d2449b7fe4cd19cdb88ad38f8c86acd4e65d1128b874c8de6719a7778
Frequently asked questions
Which Kanana release is current?
Current releases are Kanana 2 3B Instruct, Kanana 2 1.3B Instruct, Kanana 2 30B-A3B Thinking 2601, evaluated within their respective tracks.
Can Kanana be self-hosted?
Only releases with a verified official checkpoint and supported runtime configuration are presented as deployable.
Which GPU should run Kanana?
GPU selection depends on the exact variant, checkpoint precision, context length and concurrency; use the release hardware recommendations rather than the family name alone.
Official references
Last verified: 2026-08-26