Perplexity Finds Mixed-Model Agent Teams Beat Identical Model Copies
Perplexity reported that on Terminal-Bench 4.1 and WANDR, mixed teams of different frontier models outperformed identical model copies while costing less; CEO Arav Srinivas plans to iterate toward an automated researcher built on a multi-model, multi-agent backbone.
2026-11-18 ~ 2026-11-18 · 4 related posts
- Perplexity: mixed-model agent teams beat four copies of any single model, at lower cost — perplexity_ai · 2026-11-18
- Perplexity CEO: diverse multi-model agent teams beat homogeneous copies on WANDR and Terminal Bench — AravSrinivas · 2026-11-18
- Arav Srinivas: building a multi-model, multi-agent harness toward an automated researcher — AravSrinivas · 2026-11-18
1 near-duplicate retellings: perplexity_ai