What is live in Pearl right now

Every operator was told to upgrade to v1.5.0 (opens in a new tab) as soon as possible, miners and pools first. v1.5.2 followed on Oct 17, 2026. See the October security releases for what changed.

Together AI announced a Pearl-powered Gemma endpoint on May 15, 2026. Pearl’s network page describes it as priced “25%+ below the standard price”, funded by the PRL mined during the same computation. As of October 20, 2026, Together’s own page for the model says it is not available on Together’s serverless API, so treat the integration as unconfirmed until Together lists it again.

Pearl Research Labs’ own site now leads with an inference business: an OpenAI-compatible serverless API plus managed and on-premise deployments, serving open models from DeepSeek, Google (Gemma), Qwen and Z.ai (GLM). The site doesn’t say how much of that serving is mined on Pearl.

• The independently verified share of mining tied to outside AI workloads.
• Active users, transaction demand and production inference volume.
• Pool and provider concentration, measured with one consistent method.
• Long-term customer savings after rewards, liquidity, integration and operating costs.
• When, or whether, the FP8 scheme activates on mainnet.

Something becomes “live” here only when a primary source exposes a usable release, endpoint or public network behavior. Announcements stay announcements. Third-party measurements are credited to their authors rather than rewritten as universal facts.