Replicate
Replicate scores 78/100 on xpay's Agent-Ready Index v2: partial pricing visibility; self-serve checkout is available; a public API exists with documentation; usage-based / per-token pricing is published.
Replicate pricing
11 SKUs priced in mixed units (tokens, second)
| Model / SKU | Input | Output |
|---|---|---|
anthropic/claude-3.7-sonnet | $3 / 1M tokens | $0.015 / 1K tokens |
black-forest-labs/flux-1.1-pro | — | $0.04 / image |
black-forest-labs/flux-dev | — | $0.025 / image |
black-forest-labs/flux-schnell | — | $3 / 1K image |
deepseek-ai/deepseek-r1 | $3.75 / 1M tokens | $0.01 / 1K tokens |
ideogram-ai/ideogram-v3-quality | — | $0.09 / image |
recraft-ai/recraft-v3 | — | $0.04 / image |
wavespeedai/wan-2.1-i2v-480p per second of output video | — | $0.09 / second |
wavespeedai/wan-2.1-i2v-720p per second of output video | — | $0.25 / second |
cpu-small CPU (Small) compute | $0.000025 / second | — |
gpu-a100-large Nvidia A100 (80GB) GPU | $0.0014 / second | — |
How Replicate scores on the agent-ready dimensions
Public pricing visibility
15 / 15
Self-serve checkout
15 / 15
Public API + docs
16 / 20
Usage-based / metered pricing
20 / 20
Free entry path
0 / 7
No sales call required
7 / 7
Card / wallet / x402 payment
5 / 6
Machine-readable bonus (MCP / OpenAPI)
0 / 10
Total
78 / 100
What it actually costs at typical scale
1 hour of CPU (Small) compute
$0.09/mo
Can an autonomous agent transact here?
Can create account
Can enter payment method
Can upgrade plan
Can buy more usage
Can cancel online
Payment methods
Human approval recommended
Pricing transparency
full
Purchase motion
self_serve
Demo required
Quote required
Annual commit required
Auto-renewal mentioned
Cancellation policy visible
Refund policy visible
Invoice payment available
Purchase orders supported
MSA required
DPA available
Security review required
API & developer surface
Trust & compliance
SSO / SAML
RBAC
Audit logs
Data residency
Positioning & ecosystem
Buyer persona
Six-step check: can an agent actually buy from Replicate?
Discover price
https://replicate.com/pricing — full hardware + per-output tableSelect a plan
Public vs. private models with distinct billing visiblePay per task
Per-second/per-image/per-token billingAvoid a sales call
All tiers self-serve; only volume discounts gatedAPI docs without auth
https://replicate.com/docs publicEstimate cost upfront
Per-output cost shown on every model cardPros and cons for AI agents
Observational summary written by xpay from the signals captured on 2026-05-03. Not a review of the product — only of its current pricing posture for agent buyers.- Pricing is publicly visible on an indexable page — agents can read tiers without scraping past auth.
- Per-unit billing is published, so an agent can budget for a single task before committing.
- Every advertised tier is self-serve; no tier requires scheduling a call.
- API documentation is reachable without a login — discovery and integration can happen in one session.
- Per-unit rate is concrete enough that an agent can model expected spend before issuing a request.
- PPU requires account creation and an issued API key; a fully unauthenticated agent purchase is not yet supported.
- No /.well-known/ai-pricing.json or equivalent machine-readable pricing manifest is published — agents must rely on HTML scraping.
How Replicate could lift its score
Add a per-unit price (e.g. $X / 1K calls) to the pricing page so an agent can compute its own cost before committing.
| pricing_visible | true |
| headline_phrasing | Pricing — You only pay for what you use on Replicate. Some models are billed by hardware and time, others by input and output. |
| tier_count | 4 |
| lowest_paid_entry_usd | null |
| free_tier | false |
| free_tier_terms | null |
| per_unit_price | $0.09 / second |
| annual_required | false |
| self_serve_paid_tiers | 3 |
| sales_only_tiers | 0 |
| public_api_docs_url | https://replicate.com/docs |
| api_docs_auth_walled | false |
| ai_pricing_json_present | false |
| agents_txt_present | false |
| anonymous_purchase_path | false |
| per_unit_classification | {"unit_class":"compute_time","confidence":"low","agent_task_price_forecastable":"partial","rejected_reason":null,"evidence_text":"$0.09 / second"} |
| usage_headline_present | false |
| custom_tier_present | true |
| agent_friendly | {"ai_pricing_json":false,"agents_txt":false,"llms_txt":true,"sitemap_xml":true,"mcp_server_card":false,"agent_skills_index":true,"x402_supported":false} |
View raw extracted page text →
Claim this scorecard & lift your score.
Get the breakdown of every signal we measured, the one-line fix, and how your peers in AI Inference stack up.
Compare the agent-ready picks in AI Inference.
If you're building or running an AI agent that needs to buy from this category, see who's scoring highest right now.
See agent-ready picksReplicate
78
/ 100 (rubric v2.0)AI Inference
Not advertised
4
No
Sales-led
$0.09 / second
3 / 0
Public
No
Not published
2026-05-09
Per-model / per-SKU pricing
Replicate pricing starts at $0.000025 per second for the smallest CPU instance; no free tier; usage billed per second, per token, or per image depending on the model.
CPU (Small)
Discovery files and protocols
Side-channel signals — informational, not part of the score. Each protocol is independent; adoption signals the publisher is thinking about agent buyers.ai-pricing.json
agents.txt
llms.txt
sitemap.xml
MCP Server Card
Agent Skills
x402 / MPP
Compare Replicate with peers
Highest-scoring companies in the same category that an AI agent might evaluate as an alternative.Groq
87
OpenAI
81
RunPod
79
Anthropic
76
Modal
76
DeepInfra
76
Together AI
74
Fireworks AI
74
Explore the Agent-Ready SaaS Index
25,480+ SaaS scored on agent-buyability — browse by readiness band, category, or jump straight to a pay-per-run tool on xpay.tools.