Together AI and Pearl Research Labs Team Up to Reduce the Cost of AI Inference

2026-05-15 · Source: Together AI | The AI Native Cloud - Together.ai · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Blockchain & Distributed Ledger Technology · Depth: Intermediate, quick

Summary

Together AI, aiming to make AI widely accessible, announced an exclusive partnership with Pearl Research Labs on May 15, 2026. This collaboration introduces a new inference endpoint for the "Gemma-4-31B-it-pearl" instruction-tuned model, discounted by over 25%. The discount is enabled by Pearl Research Labs' Pearl Network, a blockchain protocol that utilizes Proof of Useful Work. This protocol harnesses the computation from AI inference and training workloads to secure its blockchain and mine a new cryptocurrency, ¶PRL. Pearl's mathematical breakthrough allows generating ¶PRL by obtaining proofs of matrix multiplications during AI computations, effectively offsetting AI workload costs. Together AI plans to expand its Pearl-powered product portfolio and eventually offer customers direct access to a portion of the currency emissions, aiming to further reduce LLM price-per-token as the ¶PRL spot price increases.

Key takeaway

For AI Engineers or MLOps teams focused on optimizing large language model inference costs, consider evaluating Together AI's new "Gemma-4-31B-it-pearl" endpoint. This service offers over 25% cost reduction, achieved through the Pearl Network's Proof of Useful Work, which simultaneously mines cryptocurrency. Your team can significantly lower operational expenses for specific LLM workloads, and future offerings may provide direct access to these crypto emissions for further savings.

Key insights

AI inference and training computations can simultaneously secure a blockchain and mine cryptocurrency, directly reducing AI workload costs.

Principles

• GPU cycles for AI can also produce proof-of-work digital assets.
• Proof of Useful Work increases GPU economic throughput.
• Crypto emissions can directly reduce LLM price-per-token.

Method

The Pearl Network generates ¶PRL cryptocurrency by obtaining strong proofs of matrix multiplications occurring during the forward and backward passes of AI training and inference computations.

In practice

• Access "Gemma-4-31B-it-pearl" for over 25% discounted inference.
• Anticipate future products offering direct crypto emission access.

Topics

• Together AI
• Pearl Research Labs
• AI Inference Cost Reduction
• Proof of Useful Work
• Gemma-4-31B-it-pearl
• Blockchain Protocol

Best for: Machine Learning Engineer, NLP Engineer, CTO, MLOps Engineer, AI Engineer, Director of AI/ML

Related on AIssential

• From Gold Rush to Factory: How to Think About TCO for Enterprise AI — Enterprise AI TCO requires a "factory mindset" matching diverse workloads to appropriate, cost-effective compute resources.
• Tokenomics Is Dead. The Lie Of Affordable Inference — Current "cheap inference" is a VC and hyperscaler-subsidized illusion, masking significant AI industry losses.
• Gemini Seizes the Lead, Investors Panic Over Agentic AI, Optimism at Global AI Summit, Local Versus Cloud — AI advancements are driving model efficiency, shifting global focus to benefits, and impacting software markets.
• Cheaper AI Does the Opposite of What You'd Expect — Cheaper AI inference drives increased consumption and integration, leading to better user experiences and higher demand.
• This 11-person London startup wants to make AI 100x cheaper every year — AI inference costs can be drastically reduced through hardware and software optimization.
• How to Cut Inference Costs Without Slowing Down Your AI Stack* — Iteration-level scheduling can significantly reduce AI inference costs and improve throughput and reliability.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Together AI | The AI Native Cloud - Together.ai.