PPLX Decider Models by Perplexity Labs
Last refreshed 2026-11-13. Next refresh: weekly.
Details
Capabilities
About
PPLX Decider is Perplexity's family of open-weight multimodal decision models, served through the Perplexity Decisions API (POST https://api.perplexity.ai/v1/decisions). A decider model reads a state (text, JSON, or images) plus up to 128 named typed questions (noul for yes/no, choice, or score) and returns probabilities instead of generated text. The 27B line, fine-tuned from Qwen3.8-27B, runs pplx-decider-v1-27b (2026-10-01), pplx-decider-v1.1-27b (2026-10-06) and pplx-decider-v1.2 (2026-10-16). pplx-decider-flash-v1.2 (2026-10-16) is a 4B model fine-tuned from Qwen3.5-4B and distilled from the 27B models. Weights are on Hugging Face under Apache 2.0.
Decision facts
- Best fit
- agentsvision and multimodal workstructured outputs
- Capability starting point
- pplx-decider-v1.2 with 256K context and structured outputs and multimodal inputs
- Lowest tracked input
- pplx-decider-flash-v1.2 · $0.010/1M · Perplexity Labs
- Closest related family
- Agents-A1
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 256K context, 27B parameters, and structured outputs.
Use when the workload needs 256K context, 4B parameters, and structured outputs.
Use when the workload needs 256K context, 27B parameters, and structured outputs.
Use when the workload needs 256K context, 4B parameters, and structured outputs.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| pplx-decider-v1.3 | Use when the workload needs 256K context, 27B parameters, and structured outputs. | 2026-11 | 256K context27B parametersstructured outputs | Current |
| pplx-decider-flash-v1.3 | Use when the workload needs 256K context, 4B parameters, and structured outputs. | 2026-11 | 256K context4B parametersstructured outputs | Current |
| pplx-decider-v1.2 | Use when the workload needs 256K context, 27B parameters, and structured outputs. | 2026-10 | 256K context27B parametersstructured outputs | Current |
| pplx-decider-flash-v1.2 | Use when the workload needs 256K context, 4B parameters, and structured outputs. | 2026-10 | 256K context4B parametersstructured outputs | Current |
Release Timeline
2 release groupsReplaced By
Keep for legacy integrations; evaluate pplx-decider-v1.2 before new work.
Keep for legacy integrations; evaluate pplx-decider-v1.2 before new work.
Specifications(6 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Structured Outputs |
|---|---|---|---|---|---|---|
| pplx-decider-v1.3 | 2026-11 | 256K | 27B | Yes | Yes | Yes |
| pplx-decider-flash-v1.3 | 2026-11 | 256K | 4B | Yes | Yes | Yes |
| pplx-decider-v1.2 | 2026-10 | 256K | 27B | Yes | Yes | Yes |
| pplx-decider-flash-v1.2 | 2026-10 | 256K | 4B | Yes | Yes | Yes |
Available From(3 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| pplx-decider-flash-v1.2 | Perplexity Labs | $0.01 | — | Serverless |
| pplx-decider-flash-v1.2 | OpenRouter | $0.01 | — | Serverless |
| pplx-decider-flash-v1.2 | Vercel AI Gateway | $0.01 | — | Serverless |
| pplx-decider-flash-v1.3 | Perplexity Labs | $0.01 | — | Serverless |
| pplx-decider-flash-v1.3 | Vercel AI Gateway | $0.01 | — | Serverless |
| pplx-decider-v1.2 | Perplexity Labs | $0.02 | — | Serverless |
| pplx-decider-v1.2 | OpenRouter | $0.02 | — | Serverless |
| pplx-decider-v1.2 | Vercel AI Gateway | $0.02 | — | Serverless |
| pplx-decider-v1.3 | Perplexity Labs | $0.02 | — | Serverless |
| pplx-decider-v1.3 | Vercel AI Gateway | $0.02 | — | Serverless |