Key Info

Cline has added Pareto 26.10 Preview, a new model that routes queries across frontier and open models, grades each answer, and returns the best one while preserving prompt cache and keeping costs low.

Highlights

  • Achieves the same DeepSWE score as Fable at $0.24 per task vs $13.41, roughly 56x cheaper
  • Routes to multiple frontier and open models and grades the responses
  • Preserves prompt cache to keep costs down
  • Available in the Cline CLI and the new Desktop app