IBM launches new AI models for businesses as Gen AI competition heats up

IBM has announced the release of its latest family of AI models, which promise a strong balance of power and cost efficiency for businesses.

The company says its Granite 3.0 models offer smaller, more business-focused use cases, unlike many existing large-scale general-purpose language models already available.

More broadly, IBM also announced other AI enhancements, including its latest Watsonx Code Assistant, powered by Granite code models, which supports languages including C, C++, Go, Java and Python.

IBM Granite 3.0 models

IBM envisions the power of Granite 3.0 by combining what are essentially reasonably small models with business data to unlock task-specific performance that rivals larger models, without the associated cost. The company stated that, through a series of tests, Granite 3.0 proved to be 3x-23 times cheaper than the big frontier models, which it could not name.

The company claimed that its 8B Instruct model, the larger of the two new 8B and 2B models, was able to keep up with similarly sized options from the Meta and Mistral in the benchmarks set by the OpenLLM leaderboard by Hugging Face.

At the same time, IBM also announced updates to its Granite Guardian 3.0 to allow developers to implement security guardrails by checking user prompts and LLM responses. This includes checking for things like social bias, hate, toxicity, profanity, violence and jailbreaking. Another test confirmed that IBM’s solution gave Granite Guardian 3.0 8B improved damage detection accuracy over all three generations of Meta’s Flame Guard models.

All Granite 3.0 models are available for download from HuggingFace, with language models 8B and 2B and Granite Guardian 3.0 models 8B and 3B also available for commercial use through IBM’s watsonx.

road Reuters

Related Posts

“It’s actually okay to miss someone”

A $37.2 million grant goes to fight addiction in KY

Victor Wanyama, wife Teshna sign hilarious contract and keep Jersey ahead of 2nd Born arrival