Skip to main content

Model Rates

Use this page to check what each model or system service costs before you deploy. Knowing the base rates will help you make the most of your credits whether you're running agents for development, production, or scale.

Good to know

If you're using SmythOS-recommended models, you don't need API keys. We handle setup, cost tracking, and rate scaling based on your plan.

Core Billing Rules

  • Your plan includes free credits every month.
  • Usage is billed pay-as-you-go after your credits run out.
  • All charges from external AI providers are non-refundable.
  • You can pause usage when a cap is reached.
Need to monitor usage?

Use the Analytics Dashboard to check daily agent activity, track token consumption, and export CSV reports.

Model Pricing

Want full plan comparisons?

This page lists per-unit costs. For limits, included features, or seat counts, jump to the Plans and Pricing section or visit the Compare Plans and Features page.

OpenAI Models

ModelInput*Cached Input*Output*Context WindowImage Handling
GPT-4o$2.50$1.25$10.00128k tokensInput only (token-based)**
GPT-4o mini$0.15$0.075$0.60128k tokensInput only (token-based)**
GPT-4.1$2.00$0.50$8.001M tokensInput only (token-based)**
GPT-4.1 mini$0.40$0.10$1.601M tokensInput only (token-based)**
GPT-4.1 nano$0.10$0.025$0.401M tokensInput only (token-based)**
o3$1.10$0.55$4.40200k tokensInput only (token-based)**
o4-mini$0.10$0.05$0.30200k tokensInput only (token-based)**

Google Gemini

ModelSearchInput*Output*ContextMedia Processing
Gemini 2.5 Flash$0.035$0.10$0.401M0.7/image, 0.1/media
Gemini 2.5 PRO ≤200k$0.035$1.25$10.001Msame as above
Gemini 2.5 PRO >200k$0.035$2.50$15.001Msame as above

Anthropic Claude

ModelInput*Output*ContextNotes
Claude 3.5 Haiku$0.80$4.00200kCost-efficient
Claude 3.7 Sonnet$3.00$15.00200kBest balance of power & price
Claude Sonnet 4$3.00$15.00200kHigh-performance
Claude Opus 4$15.00$75.00200kMost capable Claude model

Google AI Imagen (Image Generation)

ModelDescriptionVariantPrice per request
Imagen 4 UltraFor advanced use-cases or when the best image quality is neededimagen-4-ultra-generate-preview-06-06$0.060
Imagen 4Highly detailed images with strong text rendering and lighting improvementsimagen-4-generate-preview-06-06$0.040

Perplexity Models

ModelSearchInput*Output*ContextNotes
Sonar$0.005$1.00$1.00128k-
Sonar PRO$0.005$3.00$15.00200k-
Sonar Reasoning Pro$0.005$2.00$8.00128k-
Sonar Deep Research$0.005$2.00$8.00128k+$3.00 (reasoning)
Token billing explained

*All prices are per 1M tokens.
**For image input, tokens = width × height / 750.

SmythOS Core Services

ServiceCost
Task Execution$0.001 per task
Weaver Engine$3.00/1M input tokens, $12/1M output tokens
NodeJS Runtime$0.0001 per second
Vanilla JSFree

What counts as a task?

Each successful component execution in a deployed agent equals one task. No charges apply for failed runs or dev environment tests.

ExampleTasks Counted
3 components in chat3 tasks
100 rows, 2 components each (bulk run)200 tasks
Failed run0 tasks
Testing in builder0 tasks
Production only

Tasks are only counted when your agent is deployed. The development version is always free to run and test.

Tools and Add-ons

Image Generation

ToolProviderCost
GPT ImageOpenAI$5 input, $10 image input, $40 output / 1M tokens
HD 8 StepsFlux Schnell$0.001 per request
HD 50 StepsFlux Dev$0.008 per request

Web Utilities

ToolProviderCost
Web SearchTavily$0.008 per search
Web ScrapingScrapFly$0.00015/credit

System-Level Components

ComponentCost
NodeJS$0.0001 per second
Computer Use$0.0025 per second
MCPVaries by model

Plans and Pricing

Need more feature comparisons?

You can see what each plan includes on the Compare Plans and Features page.

Each plan includes credits, usage multipliers, and seat limits. As you grow, your rates improve.

PlanPriceSeatsCreditsMultiplierFeatures
Free Forever$01$52.5xLocal deploy, public agents
Builder$39/seat/mo1$201.5xPrivate agents, 100 API calls/day
Startup$399/mo3$2001.3xRAG support, 5,000 API calls/day
Scaleup$1,499/mo5$3001.1xWhite labeling, 25,000 API calls/day
EnterpriseCustomCustomCustomCustomUnlimited access and custom deployment
About RAG components

RAG tools are only available on the Startup plan or higher. If you're on the Builder plan and need retrieval, consider upgrading.

Avoid surprise charges

Track daily usage from the Analytics page and manage spending caps under My Plan.

What's Next?