India has nearly 1.4 billion people and is one of the most linguistically diverse countries in the world, with more than 1,500 languages and dialects spoken across different regions. For AI systems and large language models (LLMs), this creates both a major opportunity and a significant challenge, especially because high-quality training data for many Indic languages remains limited. Nevertheless, tech giant IBM (IBM) has designed its Granite 4.0 language models for these types of challenges.
TipRanks Black Friday Sale
- Claim 60% off TipRanks Premium for the data-backed insights and research tools you need to invest with confidence.
- Subscribe to TipRanks' Smart Investor Picks and see our data in action through our high-performing model portfolio - now also 60% off
In fact, they use a new hybrid Mamba-transformer architecture that reduces memory needs and allows the models to run on cheaper GPUs without losing performance. The models are open-sourced under the Apache 2.0 license and are the first open models to receive ISO 42001 certification, with cryptographic signatures for security and transparency. Interestingly, Granite 4.0 was trained on huge amounts of Indian-language data (about 100 billion tokens) along with 1.5 million post-training examples that include translated English datasets and AI-generated conversations.
This large and carefully filtered training set helps the models handle Indian languages better than earlier versions. Tests also show that Granite 4.0 performs very well compared to other multilingual models like Llama and Gemma. Moreover, smaller Granite versions, such as Granite-4.0-h-tiny, perform strongly because of their efficient Mixture-of-Experts design, while the larger 30B Granite-4.0-h-small model leads its size category against almost all competitors.
Is IBM a Buy, Sell, or Hold?
Turning to Wall Street, analysts have a Moderate Buy consensus rating on IBM stock based on eight Buys, five Holds, and one Sell assigned in the past three months, as indicated by the graphic below. Furthermore, the average IBM price target of $300.58 per share implies that shares are trading near fair value.
