- Services
Perplexity Decider v1.3 is a 27B open-weight decision model built on Qwen3.8-27B, released in November 2026. It answers several classification questions about one shared input in a single pass, returning probabilities for yes/no, multiple choice (up to 255 options) and ordered scores, plus a new locate mode that pulls verbatim text spans with character offsets and confidence scores. Weights are on Hugging Face under Apache 2.0, and the model is also served through OpenRouter and the Perplexity Decisions API at $0.02 per million input tokens with output tokens free. It accepts text and images and is aimed at routing, moderation, reranking, tagging and judging model outputs.
A workable 27B-parameter dense decision model from Perplexity. Treat the modality benchmarks above as the leading indicator of fit — composite scoring across modalities is still maturing. Newly released, so production-readiness is still being shaken out.
Generated from this model’s benchmarks and ranking signals. Editor reviews refine it over time.
Access model weights, configuration files, and documentation.
No benchmark data available for this model yet.
Benchmark data sourced from: Hugging Face, Arena.ai, Artificial Analysis.
The top devices for this model at 4-bit, ranked by fit and speed.
| Device | Grade | VRAM |
|---|---|---|
| Acer Veriton GN100 AI MiniAcer | SS | 17.0 GB |
| AI TOP ATOM ATAGB10-9003Gigabyte | SS | 17.0 GB |
| AMD Instinct MI300XAMD | SS | 17.0 GB |
| AMD Instinct MI325XAMD | SS | 17.0 GB |
| AMD Instinct MI350PAMD | SS | 17.0 GB |
Cheapest current cloud rentals with at least 17 GB VRAM, refreshed hourly.
Advertising disclosure: we earn commissions when you shop through the links below.
| Option | Cost / GPU-hour |
|---|---|
NVIDIA RTX PRO 4000 BlackwellVast.ai · Spot · 24 GB VRAM | $0.11 |
NVIDIA GeForce RTX 3090Vast.ai · Spot · 24 GB VRAM | $0.11 |
NVIDIA GeForce RTX 3090Vast.ai · On-Demand · 24 GB VRAM | $0.11 |
NVIDIA RTX A5000RunPod · Community · 24 GB VRAM | $0.16 |
NVIDIA GeForce RTX 4090Vast.ai · Spot · 24 GB VRAM | $0.20 |
Per-GPU rate across RunPod, the Vast.ai marketplace, DigitalOcean, and Vultr.
Spot tier is interruptible. Plan for restarts when comparing against on-demand prices.
Perplexity Decider v1.3 is the successor to Decider v1.2. A request consists of a shared state plus one or more questions; each question is isolated from the others and answered in one forward pass, avoiding token by token generation. Question types are noul (P(true) for yes/no), choice (1 to 255 custom options with per-option probabilities and confidence), score (ordered levels) and locate (verbatim spans from a string state with character offsets and a score). It supports text and image input, though locate reads text only and does not work with web grounding. Running the open weights needs Python 3.12+ and a CUDA GPU with room for about 49 GiB of BF16 weights plus working memory; the local default is 32,768 tokens while the API accepts under 256K tokens. A smaller 4B sibling, Decider Flash v1.3, is distilled from this model.
Explore the Provider
Aggregate stats, leaderboard, release timeline, and benchmark coverage across every Perplexity model we track.
Check which of your devices can run this model and how fast.
Compare hosted per-token prices before you commit to local hardware.
Break-even math for self-hosting, renting a GPU, or paying per token.
Live status for the hosted APIs you might use instead of running this locally.
View on Hugging Face