AI Resources
InkubaLM vs Sunflower-14B
A verified, side-by-side comparison. Both records are status-checked by Findra, so you are comparing what each actually offers today, not a stale listing.
Category
AI Resources
AI Resources
Type
LLM
LLM
Country
🇿🇦 South Africa
🇺🇬 Uganda
Docs status
Docs live
Docs live
Licensing
Pricing
Free / open weights (CC BY-NC 4.0)
Open weights (Apache 2.0)
Verified
Verified
Unverified
Last verified
5 Jul 2026
5 Jul 2026
Tags
nlp, llm, african-languages, open-weights, low-resource
llm, translation, ugandan-languages, luganda, qwen3
Summary
InkubaLM-0.4B is a 400M-parameter open-weights small language model built from scratch by Lelapa AI for five low-resource African languages (isiZulu, Yoruba, Swahili, isiXhosa, Hausa, plus English/French), using a LLaMA-style architecture trained on 2.4B tokens.
Sunflower-14B is a Qwen3-14B based multilingual LLM by Sunbird AI supporting translation and text generation across 31 Ugandan languages plus English. It achieves the highest translation accuracy in 24 of 31 evaluated language pairs and targets healthcare, agriculture, education and government use cases. It is released under Apache 2.0 with quantized GGUF and W8A8 variants available.