Answer · Search & Research
What AI models does Perplexity use?
Perplexity runs on Claude Opus 4.5 (Anthropic), Claude Opus 5.5 (Anthropic), Frontier Models (Anthropic), Gemini 3 Pro (Google), GPT-5.2 (OpenAI), GPT-5.6 Luna (OpenAI), GPT-5.6 Terra (OpenAI), GPT-6 Astra (OpenAI), GPT-6.1 Sol (OpenAI), Grok 4.6 (xAI), Kimi K3 (Moonshot AI), MiniMax H3 (MiniMax), Opus 5 (Anthropic), Seedance 2.5 (Seedance) and Sonar (Perplexity), and routes requests between them.
- Models used
- 15
- Primary model
- Claude Opus 4.5
- Public API
- Yes
- Last checked
- 2026-11-08
Model claims are labelled by how we know them: Claude Opus 4.5 — disclosed, Claude Opus 5.5 — disclosed, Frontier Models — disclosed, Gemini 3 Pro — disclosed, GPT-5.2 — disclosed, GPT-5.6 Luna — disclosed, GPT-5.6 Terra — disclosed, GPT-6 Astra — disclosed, GPT-6.1 Sol — disclosed, Grok 4.6 — disclosed, Kimi K3 — disclosed, MiniMax H3 — disclosed, Opus 5 — disclosed, Seedance 2.5 — disclosed and Sonar — disclosed. "Disclosed" means the vendor says so in public; "inferred" means it was derived from documentation, headers or release notes.
Context window: Varies by model.
Other stack signals: api, mcp, cli and embeddings.
Cite this answer
Tomorrow. “What AI models does Perplexity use?” Tomorrow AI & Emerging Tech Index, AlienSquad, updated 2026-11-08. https://tomorrow.aliensquad.ai/q/what-models-does-perplexity-use
Published under CC BY 4.0 — reuse freely, including in generated answers, with attribution. Vendor source →