perplexity
Perplexity: Decider V1.2
perplexity/pplx-decider-v1.2
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
Perplexity's Decider V1.2 is a 27B-parameter decision model fine-tuned from Qwen3.8-27B that, instead of generating text, reads a state plus structured questions and returns choices, probabilities, or scores. Its documentation describes a joint request format in which one sequence carries several questions at roughly the cost of one. The OpenRouter catalog lists text and image inputs and a 262,144-token context, though training used about 32k. On Perplexity's own Decision Index suite it scores higher than the prior checkpoint; independent verification was not available in this pack.[1][2]
Strengths
- Per its model card, a joint request format lets one sequence carry several questions about the same state at roughly the cost of answering one.[2]
- Accepts text and images up to about 1.6 MP per input, according to the model card and the OpenRouter catalog listing.[1][2]
- Perplexity's reported measurements show gains over the previous checkpoint on its text and vision decision suites, and the checkpoint ships a fitted probability-calibration temperature.[2]
Limitations
- The model card warns that grouping independent or competing candidates in one request can reduce accuracy, and advises separate requests in that case.[2]
- Self-hosting requires a CUDA GPU with room for about 52 GiB of BF16 weights plus working memory, and the Hugging Face repository is private and requires a token, per the documentation.[2]
- Training used sequences up to about 32k tokens, so longer inputs are only spot-checked; overall, evidence is a vendor model card plus an OpenRouter listing with no independent evaluation, so validate on your own workload.[1][2]
Where to use it
- Suggested for routing or triage, such as assigning a request to a team while flagging urgency or sentiment, matching the model card's worked example.[2]
- Suggested for multi-label or facet-style decisions about a single item, a case where the model card reports that grouping questions in one request helps.[2]
- Suggested for scoring or reranking related candidates in one request, with the caveat that the source reports weaker results for independent candidates.[2]
Sources
- [1] OpenRouter catalog · Perplexity: Decider V1.2Sources checked: Nov 13, 2026
- [2] Model documentation · Perplexity: Decider V1.2Sources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
Decider V1.2 is a new 27B checkpoint of Perplexity's decision model, succeeding [Decider V1.1 27B](/perplexity/pplx-decider-v1.1-27b) with the same API contract. Instead of generating text, it reads content passed as `state`...
Specifications
- Model ID
- perplexity/pplx-decider-v1.2
- Canonical ID
- perplexity/pplx-decider-v1.2-20261016
- Creator
- perplexity
- Added to OpenRouter
- 16.10.2026
- Context length
- 262,144
- Modality
- text+image->decisions
- Input modalities
- text, image
- Output modalities
- decisions
- Token pricing
- No per-token price in the source snapshot — see OpenRouter for the billing unit
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | No measurement | — |
| Coding | No measurement | — |
| Agents | No measurement | — |
| Writing | No measurement | — |
| Reasoning | No measurement | — |
| Document analysis | No measurement | — |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
No evaluations are supplied for this model.
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| Perplexity | 262,144 | — | — | — | — | 100% |
Provider data retrieved: 13.11.2026
Related articles
No published articles identify this model yet.