Perplexity Releases Decider v1.3 and Decider Flash v1.3 Decision Models

Perplexity released Decider v1.3 and Decider Flash v1.3, scoring 70 and 64 on the Artificial Analysis Decision Index v2.1. Decider v1.3 holds the #6 position while Decider Flash v1.3 rises to #10, with both models on the Capability vs. Cost Pareto frontier at $68 and $34 per 1M tasks. Both offer a 256K-token context window, text and image input, and Apache 2.0 licensing via Perplexity's first-party API.
Event version 1.
Key insights
Decider v1.3 holds #6 ranking
Decider v1.3 scores 70 on the Artificial Analysis Decision Index v2.1 and maintains the #6 position held by its v1.2 predecessor. It ranks #4 on AA-Paperclip with a score of 69, a 2-point improvement, and #3 on EvalSafe-O*NET-AA with 89, a 1-point gain.
Decider Flash v1.3 climbs to #10
Decider Flash v1.3 scores 64 on the Artificial Analysis Decision Index v2.1, a 4-point improvement over v1.2, and moves up into the #10 position. The gain is driven by a 12-point improvement on both AA-Symmetry and When2Call.
Both models on Capability vs. Cost Pareto frontier
Decider v1.3 and Decider Flash v1.3 occupy positions on the Capability vs. Cost Pareto frontier. Cost per 1M tasks is $68 for Decider v1.3 and $34 for Decider Flash v1.3.
Shared specifications and availability
Both models offer a 256K-token context window, text and image input, and Apache License 2.0. Input pricing is $0.02 per 1M tokens for Decider v1.3 and $0.01 for Decider Flash v1.3, with sizes of 27B and 4B respectively, available through Perplexity's first-party API.