OpenAI tests outcome-based billing, plus two more AI shifts

Today's AI Newsroom covers outcome-based AI pricing, Anthropic's latest compute agreement, and new research into smaller rubric-based reinforcement learning judges.

OpenAI tests outcome-based billing

OpenAI has begun letting some of its largest customers pay only when its AI actually completes the job. The arrangement is limited to select major accounts rather than offered generally.

• Intercom charges $0.99 for each conversation its Fin agent resolves and nothing for ones it does not.
• Zendesk restricted billing to Verified Resolutions, confirmed by an LLM evaluation within 72 hours of the conversation.
• Salesforce launched Agentforce at $2 per conversation, charged for every 24-hour session whether or not anything was resolved.

Sources: The Information, The Next Web, and PYMNTS.

Anthropic signs a $35B compute deal with Lambda

Anthropic signed a $35 billion deal for computing with Lambda. The Texas facility linked to the project in Nueces County is being built by Hut 8.

• Lambda is a cloud company backed by Nvidia.
• Nvidia is expected to be the facility's lessee.
• Lambda has been negotiating a fundraising of up to $3 billion.

Sources: Moomoo, Briefs, and Yahoo Finance.

Study tests small models as rubric-based RL judges

Reinforcement learning from human feedback has become the dominant paradigm for aligning large language models with human preferences. Traditional RLHF relies on scalar reward signals that lack interpretability and fail to capture the multifaceted nature of response quality.

• Rubric-guided reinforcement learning introduces structured, interpretable evaluation criteria as its backbone.
• Rubric-based reinforcement learning extends RL beyond tasks with exact answers or rule-based verifiers by scoring responses against instance-specific criteria.
• Training requires repeated rubric judging, often with proprietary APIs or local generative LLM judges with 7B parameters or more.

Sources: Small Language Models as Judges for Rubric-Based Reinforcement Learning and Rubric-based Reinforcement Learning.

This issue was researched and written by the Classwise AI Newsroom. Read the canonical issue or get the free daily briefing.