Liquid AI, a Massachusetts-based artificial intelligence (AI) startup, introduced its first generative AI fashions not constructed on the present transformer structure. Dubbed Liquid Basis Mannequin (LFM), the brand new structure strikes away from Generative Pre-trained Transformers (GPTs) which is the inspiration for widespread AI fashions such because the gpt sequence by OpenAI, Gemini, Copilot, and extra. The startup claims that the brand new AI fashions have been constructed from first rules and so they outperform giant language fashions (LLMs) within the comparable dimension bracket.

Liquid AI’s New Liquid Basis Fashions

The startup was co-founded by researchers on the Massachusetts Institute of Know-how (MIT)’s Pc Science and Synthetic Intelligence Laboratory (CSAIL) in 2023 and aimed to construct newer structure for AI fashions that may carry out at an identical stage or surpass the GPTs. .

These new LFMs are available in three parameter sizes of 1.3B, 3.1B, and 40.3B. The latter is a Combination of Specialists (MoE) mannequin, which suggests it’s made up of assorted smaller language fashions and is geared toward tackling extra advanced duties. The LFMs at the moment are obtainable on the corporate’s Liquid Playground, Lambda for Chat UI and API, and Perplexity Labs and can quickly be added to Cerebras Inference. Additional, the AI ​​fashions are being optimized for Nvidia, AMD, Qualcomm, Cerebras, and Apple {hardware}, the corporate stated.

LFMs additionally differ considerably from the GPT expertise. The corporate highlighted that these fashions have been constructed from first rules. The primary rules are primarily a problem-solving method the place a posh expertise is damaged right down to its fundamentals after which constructed up from there.

In accordance with the startup, these new AI fashions are constructed on one thing referred to as computational items. Put merely, this can be a redesign of the token system, and as an alternative, the corporate makes use of the time period Liquid system. These include condensed data with a give attention to maximizing data capability and reasoning. The startup claims this new design helps cut back reminiscence prices throughout inference, and will increase efficiency output throughout video, audio, textual content, time sequence, and alerts.

The corporate additional claims that the benefit of the Liquid-based AI fashions is that its structure might be mechanically optimized for a selected platform based mostly on their necessities and inference cache dimension.

Whereas the claims made by the startup are tall, their efficiency and effectivity can solely be gauged as builders and enterprises start utilizing them for his or her AI workflows. The startup didn’t reveal its supply of datasets, or any security measures added to the AI ​​fashions.