Agent & decision APIs compared: pricing, limits, status

22 APIs·Last verified 2026-10-16·Full API index·Comparisons·Estimate cost·Free no-key APIs for agents →

Featured

Verified & partly verified

NewBetaVerified

OpenRouter Armor

OpenRouter · Guardrail / tool-call policy

Armor evaluation is not billed during the beta. OpenRouter says pricing will be announced before general availability. Normal model inference is still billed.

OpenAI Decisions API

OpenAI · Decision model

$0.10 per 1M input tokens; no output, cache-read, or cache-write charges.

Microsoft-Decision-1

Microsoft (Foundry) · Decision model

$0.042 per 1M input tokens; output $0 (OpenRouter listing, provider Azure).

Vercel AI Gateway — Decisions endpoint

Vercel · Gateway / proxy

No markup and no platform fee on tokens: you pay the provider’s list price (also with your own key). Per-request cost is reported in response metadata; the docs’ example shows $0.10 per 1M for openai/gpt-6-luna-decisions, matching OpenAI’s list price.

Jev (TypeSafe AI)

TypeSafe AI · Decision model

$0.042 per 1M input tokens ($42 per billion) for both Jev 1.14 and 1.13; output tokens free. Long-context mode costs the same per token.

Cloudflare Clef (Clef, Clef-flash, Clef-omni)

Cloudflare (Workers AI) · Decision model

Clef $0.240, Clef-omni $0.150, Clef-flash $0.038 per 1M input tokens; no output-token charge. Media are converted to input tokens.

Sponsor this slot

All APIs

16 more

Perplexity Decisions API (pplx-decider)

Perplexity · Decision model

$0.02 per 1M input tokens on pplx-decider-v1.2 and $0.01 on pplx-decider-flash-v1.2; output tokens free; no per-request fee. State, questions, and images count as input. In Perplexity’s tests an image cost about 1,000 input tokens per megapixel.

Inception Mercury Decide

Inception · Decision model

Input listed as $0.04 → $0.02 per 1M tokens (shown as a 50%-off promotion); cached input and output $0.

Liquid AI · Decision model

$0.02 per 1M input tokens, $0 output, on Liquid’s pricing page. New orgs get $2 free credit on Liquid Console.

Strands Decider 2B (open source)

Strands Agents (AWS) · Open-weight decision model

Free open weights (Apache-2.0); you pay your own compute.

AWS Bedrock Guardrails

AWS · Policy guardrail

Content filters (text) $0.15 per 1,000 text units; denied topics $0.15/1k; sensitive-info filters $0.10/1k; regex and word filters free; contextual grounding $0.10/1k; Automated Reasoning checks $0.17/1k per policy; image content filter $0.00075/image. A text unit is up to 1,000 characters.

Content Safety in Foundry Control Plane (formerly Azure AI Content Safety)

Microsoft · Harm classifier

Free (F0) tier: 5,000 text records/month and 5,000 images/month; usage stops at the limit. A text record is up to 1,000 characters.

Gemini structured output (Vertex AI / Agent Platform)

Google Cloud · Structured output

Per-token Gemini pricing; for example Gemini 3.8 Flash is $0.75 / $3.75 per 1M input/output tokens (introductory, through 2026-12-31), then $1.50 / $7.50 from 2027-01-01. The response schema counts toward the input token limit.

Anthropic Claude (tool use / structured outputs)

Anthropic · Structured output

Per 1M input/output tokens: Haiku 5.5 $0.10/$0.50 (prompts up to 100k); Haiku 4.5 $1/$5; Sonnet 5.5 $2/$10; Opus 5.5 $4/$20. US-only inference_geo costs 1.1x; Batch is 50% off.

OpenAI Moderation (omni-moderation)

OpenAI · Harm classifier

Cohere Classify

Cohere · Classifier

No Classify price found. cohere.com/pricing does not list Classify (it lists Model Vault instances and legacy Command/Aya token prices). Trial keys are free but rate limited and not for production.

Laya (open source)

Convai Innovations · Open-weight decision model

Free open weights (Apache-2.0); you pay your own compute.

llama.cpp decision endpoint (/v1/systemone)

ggml-org (llama.cpp) · Open-weight decision model

Free and open source; you pay your own compute. Check each model’s license (OpenJev is CC BY-NC 4.0, non-commercial).

DeepSeek Search & Fetch API

DeepSeek · Web search & fetch (agent tool)

Search $0.75 and Fetch $0.75 per 1,000 successful requests. Only HTTP 200 responses with a result are billed: empty searches, failed fetches and timeouts are free, and count does not change the price. Charges come out of your topped-up or granted balance.

Liquid AI d1-ultra

Liquid AI · Decision model

$0.042 per 1M input tokens, cached input $0.042, output $0. Liquid says the price is the same on Liquid Console, OpenRouter and Vercel AI Gateway; new Liquid Console organizations get $2 free credit.

H2O-Lightning-4B (open source)

H2O.ai · Open-weight decision model

Free open weights (Apache-2.0); you pay your own GPU. H2O estimates $0.021 per 1,000 decisions self-hosted (its own estimate).

Partly verified

LLM Gateway — Decisions

LLM Gateway · Gateway / proxy

Jev 1.13: $0.042 per 1M input tokens, output free. Prices for the other decision models on the gateway were not checked.