Skip to content
Made with Jev

Jev vs Perplexity Decisions

Perplexity shipped a Decisions API in September 2026: model pplx-decider-v1-27b handling noul/choice/score questions, 1–128 questions per call, images via base64, $0.04/M input (output free). TypeSafe's Jev launched the same month. Both are GA decision APIs; Perplexity has production vision and slightly lower per-token cost, Jev has documented builds and, since 14 October 2026, image input and a Box answer in preview.

Updated · by

In short

  • Jev: text plus PNG/JPEG images in preview, $0.042/M input (output free), GA via TypeSafe API, closed weights, madewithjev proof builds.
  • Perplexity Decisions: multimodal (images via base64), $0.04/M input (output free), 1–128 questions, input under 262k tokens.
  • Both: GA, noul/choice/score question types, per-option probabilities and confidence, batch multiple questions per call.
  • Difference: ~5% per-token cost, vision support, proof builds. Same category validation: typed decisions beat generative prose for classify/route/act.

What Perplexity announced

From the Perplexity Decisions quickstart:

  • Model: pplx-decider-v1-27b
  • Endpoint: POST https://api.perplexity.ai/v1/decisions
  • Question types: noul, choice, score
  • Input: text and images (images via base64 data URLs in state)
  • Batch: 1–128 questions per call
  • Context: input under 262,144 tokens
  • Pricing: $0.04 per million input tokens, output free

Source: Perplexity Decisions quickstart

What Jev is

TypeSafe's Jev is a decision model (System One), not a chat model that happens to pick a label. Launched September 15, 2026.

  • Question types: Choice, Score, Noul (mixable in one call)
  • Returns: per-option probabilities and confidence
  • Input: text (string, JSON object, or array of strings), plus PNG and JPEG images in preview on 1.13 and 1.14 with the images header. No audio or video
  • Box: one bounding box on an image, Jev 1.14 only (preview)
  • API: documented POST /v1/systemone with Python and JavaScript SDKs
  • Pricing: $0.042 per million input tokens, output free
  • Batch: ~60 questions cost about the same as one
  • Limits: 64k tokens per request and 32k per question by default; 262k for both with the opt-in long-context header (preview)

Source: TypeSafe models

Side by side

Facts only. Where either product has not published a detail, that is what the table says.

As of early October 2026 reporting. Both products evolve; verify current docs before deciding.
TypeSafe JevPerplexity Decisions
AnnouncedSep 15, 2026Sep 2026
AvailabilityGenerally availableGenerally available
ModelPurpose-built System Onepplx-decider-v1-27b
Public contractDocumented POST /v1/systemone + SDKsDocumented POST /v1/decisions
Question typesChoice, Score, Noulnoul, choice, score
Probabilities / confidencePer-option probs + confidencePer-option probs + confidence
Input modalitiesText; PNG and JPEG images in previewText + images (base64)
Bounding box answerBox type, Jev 1.14 only (preview)Not documented
Max questions per call~60 cost about the same as one1–128
Context limit64k per request, 32k per question; 262k opt-in (preview)Input under 262,144 tokens
Price$0.042 / M input; output free$0.04 / M input; output free
Open weightsNo (closed)Not published (as of early Oct 2026)
Self-hostableNoNot published
Proof buildsmadewithjev catalogVendor benchmarks only

Price difference: Perplexity $0.04/M vs Jev $0.042/M input = $0.002/M tokens or ~5%. Marginal difference; both significantly cheaper than chat models for routing/classification.

Latency caveat: Neither vendor publishes P50/P99 latency as of early October 2026. Do not compare vendor slides to third-party measured figures without labeling methods.

What builders should do

  1. If you need vision (screenshots, UI state): Perplexity documents images via base64 data URLs. Jev reads PNG and JPEG in preview only; for production, preprocess images with OCR or vision models if using Jev.
  2. If you want proof builds: Jev has madewithjev catalog with published cost/latency figures. Perplexity: vendor benchmarks only as of early October 2026.
  3. If you want lowest per-token cost: Perplexity $0.04/M, Jev $0.042/M — difference is ~5%. Cloudflare's Clef-flash is lower than both at $0.038/M, with a 24k hosted context window.
  4. If you are already using Perplexity APIs: Adding Decisions to existing Perplexity billing is simpler than wiring a second vendor.
  5. If you need agent orchestration patterns: Jev has documented skill/MCP/orchestration rules for Claude Code, Cursor, Codex. Perplexity: check current docs for agent integration.

Builds that show the job

Routing, classification, and agent next action — with the cost and latency each builder published.

Duncan on X

Duncan

@ephraimduncan

Built a model router with Jev by @typesafeai. Jev decides what model fits your request best and the request is sent to that model.

Video from @ephraimduncan

XRouting and model choice

A model router on Jev

Gregor Zunic on X

Gregor Zunic

@gregpr07

Breaking: Browser Use + Jev = Ultrafast ⚡ Findings flights took 7s and cost only $0.0039 🤯 > new action space every step > DOM state space > small LLM fallback to type (this video is at 1x speed btw) Built a tiny open source browser agent. try it below ↓

Video from @gregpr07

XSearch

Pick

Flight search with Browser Use

Run time
~7 s
Cost
~$0.004
tamara on X

tamara

@tamarajtran

found the perfect use case for @typesafeai Jev: instant compaction in 2026, why is compaction still a summarization prompt? Jev can make it instant by scoring every tool call and dropping what’s irrelevant

Video from @tamarajtran

XContext and memory

Pick

Instant compaction for Claude

Hassan on X

Hassan

@nutlope

I used Jev to classify 1,018 AI research papers. The result: $0.08 total cost and 256ms median end-to-end latency per paper. The pipeline was: 1. Summarize each paper with DeepSeek V4 Flash 2. Send the title + summary + 24 possible topics to Jev 3. Use Jev to classify each paper 4. Visualize everything on http://1kpapers.com The summaries cost $3.99 on @togethercompute. The classifications cost $0.08 on @typesafeai. So for just over $4 of inference, I ended up with a pretty useful way to explore the top AI research papers from the past year. I think this is where things are heading: different models for different parts of the workflow, instead of using one model for everything. I’m running evals on the Jev classifications before replacing the current ones, but the site is already live: http://1kpapers.com

Video from @nutlope

Live siteDocuments and OCR

1kpapers

Documents
1,018 papers
Total cost
$0.08
Median latency
256 ms

More on Jev use cases.

Try it today

Next steps

Common questions

What is Perplexity Decisions API?
A decision API on model pplx-decider-v1-27b. Handles noul/choice/score question types, 1–128 questions per call, images via base64 data URLs, input under 262,144 tokens. Generally available with documented API and pricing ($0.04/M input, output free).
How is Jev different from Perplexity Decisions?
Jev reads text, plus PNG and JPEG images in preview since 14 October 2026, at $0.042/M input (output free), closed weights, madewithjev proof builds, ~60 questions cost about the same as one. Perplexity Decisions is multimodal (images via base64), $0.04/M input, 1–128 questions, no published open weights as of early October 2026.
Which is cheaper, Jev or Perplexity Decisions?
Very close: Perplexity $0.04/M input, Jev $0.042/M input (both output free). Difference is $0.002/M tokens or ~5%. Both significantly cheaper than chat models for routing/classification.
Does Perplexity Decisions support images?
Yes. Perplexity Decisions accepts images via base64 data URLs in state, per their quickstart docs. Jev reads PNG and JPEG images in preview since 14 October 2026, with the TypeSafe-Danger-Zone-Images header; TypeSafe says it is not production-ready yet, so preprocess images with OCR or a vision model for production.
Do both return probabilities and confidence?
Yes. Both return per-option probabilities and confidence scores. Your code can act on confident answers automatically and escalate uncertain ones to a person or larger model.
Which should I use for agent routing?
If you need vision in production (screenshots, UI state): Perplexity, since Jev's image input is a preview. If you want proof builds: Jev has the madewithjev catalog. On per-token cost: Perplexity $0.04/M, Jev $0.042/M — a marginal difference — and Cloudflare's Clef-flash is below both at $0.038/M. Both are GA with documented APIs.

Made with Jev is independent and not affiliated with TypeSafe AI. Every figure on this page is the one its author published, linked to where it can be checked.