From the source

Low preset updated

Model update

The low preset of the Agent API now uses the openai/gpt-5.6-luna model with minimal reasoning effort and a 32,768-token maximum output.

Frozen configurations need manual updates; dynamic requests update automatically.

From the source
The Agent API low preset now uses openai/gpt-5.6-luna with minimal reasoning effort and a 32,768-token maximum output.
docs.perplexity.ai

Other stories

Decisions API: pplx-decider-v1.1-27b replaces v1
Perplexity replaces Decisions API model and cuts input price
Perplexity· 9 OctModel update
AI Agent vs MCP? When To Use Each (& When You Need Both)
Guide to AI agents vs MCP and when to combine them
Perplexity· 9 Oct
AI guardrails: What they are and seven practical examples
A practical guide to AI guardrails and how they work
Perplexity· 8 Oct

Announced: 31 Jul 2026 12:18 UTC
Detected: 30 Aug 2026 08:37 UTC (+29d after)

ModelsGPT-5.6 Luna, openai/gpt-5.6-luna

Visit source ↗docs.perplexity.ai

Claude Haiku 5.5

Perplexity· 7 Oct

RAG vs. LLM: Definitions, Differences, and Use Cases

Explains how RAG supplements LLMs with external data for accuracy and timeliness.

Perplexity· 7 Oct

Multimodal embeddings beyond a single vector

New late-interaction embedding models with multi-vector representations and multimodal support.

Perplexity· 7 OctNew model

Low preset updated

From the source

Decisions API: pplx-decider-v1.1-27b replaces v1

AI Agent vs MCP? When To Use Each (& When You Need Both)

AI guardrails: What they are and seven practical examples

RAG vs. LLM: Definitions, Differences, and Use Cases

Multimodal embeddings beyond a single vector