Fireworks.ai, a generative AI firm that optimizes and customizes open source LLMs for business needs.

Fireworks.ai is based in California. Artificial intelligence (AI) startup offering a unique solution for enterprises. AI firms do not build large language models (LLMs) or foundational models from scratch, but rather curate open-source models and convert them into application programming interfaces (APIs) to enable businesses to deploy AI capabilities in a seamless manner. I got help. Fine-tuning reduces the scope of the AI ​​model and focuses it on a specific functionality. This allows them to reduce AI hallucinations and significantly improve model capabilities.

The AI ​​firm was co-founded by Lin Qiao, who also holds the CEO seat at the company. After serving as senior director of engineering at Meta and working with AI frameworks and platforms, Qiao and his team founded the startup in October 2022, according to his LinkedIn profile. Profile. In a ___ The conversation With TechCrunch, he explained Fireworks.ai’s business model, highlighting the fine-tuning service they provide. “These can be either off-the-shelf, open-source models or models that we tune or models that our customers can tune themselves. All three types can be served through our inference engine API,” he said. are

This puts the firm in a unique position where it is not innovating at the foundation model level, but it is bridging the gap between LLM and business-ready products that can be deployed seamlessly. . With a primary focus on building APIs, Fireworks.ai lets its enterprise clients plug and play any open source AI model in its catalog. According to the report, the company allows entrepreneurs to experiment with different AI models to suit their needs.

At that time, the Start Claiming to contain 89 open source LLMs such as Mixtral MoE 8x7B Instructables, of meta Llama 2 70B Chat, Google’s Gemma 7B Instruct, Stability AI’s Stable Diffusion XL, and more. The AI ​​firm offers models in either a serverless format that doesn’t require businesses to configure hardware or deploy models, or on-demand models that are available for dedicated deployments, depending on business needs. are offered on specific GPU configurations.

For the on-demand format, Fireworks.ai has three payment plans — Developer, Business, and Enterprise — where the Developer plan comes with a pay-per-use structure and a rate limit of 600 requests per minute, in the Enterprise tier. As needed. Pricing offers and unlimited rate limits. The serverless format is billed on a per-token pricing plan where different models, depending on whether they are text-only, image-only, or multimodal, will fetch a different price.