LLMs Evaluating Services

The buyer seeks an onshore USA organization to provide LLM evaluation services offsite. The selected vendor must design and implement a full benchmarking ecosystem tailored to real actuarial and insurance challenges, culminating in full publication to the agency's GitHub repository. The required deliverables comprise a public-facing model comparison platform built for repeatable re-testing, a benchmark suite containing datasets, task descriptions, and scoring scripts, a model evaluation report analyzing performance on perception and classification tasks, a 1-to-2 page non-technical executive summary, replication documentation, and maintenance recommendations. The buyer requests proposals to establish the budget and accepts digital responses via email or online. Interested bidders must submit all questions no later than September 14, 2026.