Jev pricing and cost per decision model request

Jev costs $0.042 per million input tokens, with free output.

Pricing questions

How much does Jev cost?

Jev: Published price (vendor-reported): $0.042 per million input tokens; output free

Jev facts · Source: typesafe.ai ↗

Which decision models offer free API access or a free trial?

Mercury Decide: Vendor-reported facts: Price: $0.02 per million input tokens at a 50% discount, output free; a free tier remains (OpenRouter, as of 22 Oct 2026) Parallel Decisions API: Vendor-reported facts: Access: Hosted API, with a playground; 5,000 free requests to start; Decisions Fast also on OpenRouter from 6 Nov 2026, with a seven-day free promotion LingDT-Decision-1.0: Vendor-reported facts: Price: Ant lists ¥0 per million input and output tokens, with an original input rate of ¥0.20 per million tokens. Checked 04 Nov 2026. The rate is for the 4B test endpoint during a free trial; no trial end date was established. MiniCPM-OCR Jev: Vendor-reported facts: Access: Open weights in BF16, AWQ, GPTQ and GGUF; a free API on ModelBest's platform

Mercury Decide facts · Parallel Decisions API facts · LingDT-Decision-1.0 facts · MiniCPM-OCR Jev facts · Source: openrouter.ai ↗ · Source: parallel.ai ↗ · Source: maas.antdigital.com ↗ · Source: huggingface.co ↗

Is Jev output billed?

Jev: Published price (vendor-reported): $0.042 per million input tokens; output free

Jev facts · Source: typesafe.ai ↗

How much does Decider 1 cost?

Decider 1: Published price (vendor-reported): $0.03 per million input tokens; output $0

Decider 1 facts · Source: meragpt.com ↗

How much does Celeris-1-decision cost?

Celeris-1-decision: Published price (vendor-reported): $0.04 per million input tokens; output free

Celeris-1-decision facts · Source: celeris.ai ↗

Cost per decision: work from billed tokens

For an input-token API, multiply the billed input tokens by the price per million, then divide by 1,000,000. Free output removes one charge; it does not make the input free. Count the question and options as well as the state.

Source: openrouter.ai · Source: developers.openai.com

A hypothetical 1,000-decision text workload

Assume 1,000 decisions, each billed for exactly 1,000 input tokens, with no search, cache discount, retries or extra calls. That is 1,000,000 billed input tokens. These are arithmetic examples at vendor-reported rates, not usage receipts or a comparison of answer quality.

  • Jev: $0.042 for the workload, or $0.000042 per decision.
  • Kev 4B on OpenRouter: $0.042 for the workload, or $0.000042 per decision.
  • d1 or d1.1 on OpenRouter: $0.02 for the workload, or $0.00002 per decision.
  • OpenAI Decisions API: $0.10 for the workload, or $0.0001 per decision, using the recorded uncached rate.

Jev facts · Kev facts · d1 facts · OpenAI Decisions API facts

Source: openrouter.ai · Source: developers.openai.com

Per-call prices buy a different service

Parallel Decisions API is recorded at $1 per 1,000 calls and searches the live web. If each decision takes one call, 1,000 calls cost $1. Its call-wide citations and search access differ from scoring supplied context. A token price and a call price alone cannot tell you which service answers your task well.

Parallel search, citations and limits · Evaluate the same task

Source: parallel.ai · Source: docs.parallel.ai

Check how questions are billed

Liquid's docs say each question is billed as its own prompt, including the question text and all images. A shared state in your request does not by itself establish shared-state billing. Read the provider's usage fields before applying the hypothetical 1,000-token allowance to your app.

Liquid d1 family

Source: docs.liquid.ai

Self-hosting needs a cost record too

Downloadable weights move inference onto hardware you operate. The registry records hardware reports and licence terms, but has no common measurement of utilisation, electricity or operator time. It cannot give a universal buy-versus-host break-even. Record those costs alongside billed tokens and failed or escalated decisions.

Local runtimes and reported hardware · Compare model access

Source: huggingface.co · Source: github.com

A camera example: idle hours change the price

As a hypothetical hosting-cost calculation using the camera report's prices, an always-on g6.xlarge at $0.805 an hour costs about $0.000224 per second ($0.805 divided by 3,600). If it could sustain roughly 2.8 frames per second, its hosting cost per frame would match Gemma 4 on Bedrock at the reported $0.000080 per frame. The AWS article did not measure L4 throughput. Capacity at that rate and actual spending remain unverified.

Strands model and image path · Camera evidence and limits · Task-specific evaluation

Evidence: AWS Builder Center · security-camera comparison · discovered 10 Nov 2026 · Source ↗

A filter saves only the calls it removes

The report replays stored Gemma answers with a Strands filter. Its 0.1 cutoff saves 7% of estimated LLM cost with no accuracy loss on that set. At the reported prices, a hypothetical hosting-cost break-even would require sustained throughput of about 40 frames per second, which the article did not measure. The GPU still handles every frame. Sample mix, image size, spare hardware and required accuracy change that estimate; it excludes operator time and does not establish an OpenAI Decisions API result.

Prices are the rates recorded on each model's page on decisionmodels.ai. Page built 12 Nov 2026.

Compare decision models · What is a decision model? · Jev alternatives

Open-weight models are free to download; you pay for your own hardware. Hosted prices are shown where listed.

Published prices

ModelMakerGroupPriceAccess / Licence
JevTypeSafe AIhosted$0.042 per million input tokens; output freeAccess: Launched in early access; also via OpenRouter
Decisions APIOpenAIhosted$0.10 per million uncached input tokens; output freeAccess: Generally available since 22 Oct 2026; public beta launched 6 Oct; native GPT-Live integration since 15 Oct
d1Liquid AIhostedOpenRouter lists d1 and d1.1 at $0.02 and d1-ultra at $0.042 per million input tokens, output free (as of 28 Oct 2026)Open weights: d1-3B and d1-omni-600M (7 Oct), then d1-2B and d1-Omni-2B (announced 27 Oct), under the LFM Open License v1.0
Mercury DecideInceptionhosted$0.02 per million input tokens at a 50% discount, output free; a free tier remains (OpenRouter, as of 22 Oct 2026)Access: OpenRouter
Solar DecideUpstagehostedOpenRouter lists $0.05 per million input tokens, output $0 (as of 2 Oct 2026); the launch post described a limited-time 50% discountAccess: OpenRouter
Span-01Respanhosted$0.02 per million input tokens; output free
Decider 1meraGPThosted$0.03 per million input tokens; output $0
Microsoft-Decision-1Microsofthosted$0.042 per million input tokens; output freeAccess: Microsoft Foundry, OpenRouter
Clef-omniCloudflarehosted$0.150 per million input tokensAccess: Cloudflare Workers AI; Clef-flash weights are on Hugging Face
Celeris-1-decisionCelerishosted$0.04 per million input tokens; output free
SubQ SnapSubquadratichosted$0.04 per million input tokens; output freeAccess: Subquadratic's API; the Snap 1.1 Baseten listing is a separate route
ClairicCloudflarehosted$0.10 per million input tokens on OpenRouter; output freeAccess: Workers AI as @cf/cloudflare/clairic; back on OpenRouter as cloudflare/clairic when checked 4 Nov 2026
Parallel Decisions APIParallelhosted$1 per 1,000 calls (as of 25 Oct 2026)Access: Hosted API, with a playground; 5,000 free requests to start; Decisions Fast also on OpenRouter from 6 Nov 2026, with a seven-day free promotion
KevJared Palmeropen$0.042 per million input tokens and $0 per million output tokens for jaredpalmer/kev-4b on OpenRouter (VENDOR-REPORTED, 2 Nov 2026)Licence: Apache-2.0
DeciderPerplexityopen$0.02 (V1.2) and $0.01 (Flash) per million input tokens on OpenRouter, output free (as of 22 Oct 2026)Licence: Apache-2.0
Drex 2.0Nace.AIopen$0.03 per million input tokens for Drex 2.0, Drex Max, Max v1.1 and Max Finance, hosted at NaceLicence: Drex Max weights: Nace.AI Open RAIL-M · Access: Drex Max supports text and images through Python on NVIDIA GPUs. CPU and Apple Silicon runtimes support text only. Max Finance API id: drex-max-finance.
Mammoth 1Baseten (Base Labs)openOpenRouter lists mammoth-1 at $0.10, elephant-1 at $0.035 and joey-1 at $0.01 per million input tokens, output free (as of 30 Oct 2026)Licence: MIT
Hume Audio Decisions APIHume AIhosted$0.47 per million input tokens (audio, questions and context); answers not billed. Hume's billing docs, checked 03 Nov 2026Access: Playground; request API access through Hume's waitlist
Zoom Decisions APIZoomhosted$0.042 per million tokens for Decisions API; $0.10 for Pro, per Zoom's product page checked 03 Nov 2026Access: Zoom AI Services portal and playground; sign-up or login required
LingDT-Decision-1.0Ant Digital Technologies (蚂蚁数科)hostedAnt lists ¥0 per million input and output tokens, with an original input rate of ¥0.20 per million tokens. Checked 04 Nov 2026. The rate is for the 4B test endpoint during a free trial; no trial end date was established.
hs-decide-1-flashVectorize (Hindsight)open$0.02 per million input tokens on Hindsight Inference; input onlyLicence: Apache-2.0 · Access: Open weights; vLLM plugin or Transformers on CPU, Apple Silicon and supported GPUs; hosted POST /v1/systemone
ReflexMorphhosted$0.042 per million uncached input tokens; $0.008 cached; output freeAccess: POST /v1/systemone on Morph; morph/reflex on OpenRouter
GraysonFinichosted$0.035 per million input tokens for ask and score; creating and calibrating scoring models is freeAccess: Hosted POST /v1/ask with model grayson-1; no public weights confirmed
ZuseBinerichosted$0.12 per million input tokens for 9B; $0.75 for 27B; no output tokensAccess: Bineric API; dedicated and on-premises deployments by arrangement
Sarvam PrismSarvam AIopenSarvam API: ₹4 per million input tokens; output freeLicence: Apache-2.0 · Access: Public weights and prism.py; hosted POST /v2/systemone
Gemini Decision 1Google DeepMindhostedGoogle reports $0.10 per million input tokens; output free. Shared input billed once, plus each question's instructions and criteriaAccess: Gemini API in Google AI Studio, including a no-cost prototyping tier; listed on OpenRouter

Prices not recorded here

Hosted

Open weights

Tooling

Community