Skip to content
Last updated

Freddy pricing is based on synapses (output generation) and neurons (input processing) consumed by each request. This lets you pay precisely for what you use.

What Are Synapses and Neurons?

  • Neurons — Units of input processing. Every token in your messages, thread history, system prompt, and file context consumes neurons.
  • Synapses — Units of output generation. Every token the model generates in its response consumes synapses.

Read the full explanation in Synapses and Neurons.

Pricing Units

MetricRate
Neurons (input processing)8.00 CHF per 1 million neurons
Synapses (output generation)Varies by model

Pricing is billed in CHF (Swiss Francs). Your dashboard shows consumption in real time.

Model Tiers

Different models have different synapse costs:

ModelRelative CostNotes
gpt-4oStandardBalanced cost/performance
gpt-4oLowerOptimized for throughput
gpt-4oHigherAdvanced capabilities
OpenAI, Anthropic, Google modelsVariesDepends on provider pricing

Contact your account manager or check Freddy for exact per-model rates applicable to your plan.

Controlling Costs

Limit output length

Use max_output_synapses to cap how many synapses a single response can consume:

{
 "organization_id": "org_your_org_id",
 "model": "gpt-4o",
 "max_output_synapses": 512,
 "inputs": [{"role": "user", "content": "Summarize this briefly."}]
}

Common limits:

  • 512 — Short answers, single-sentence replies
  • 1024 — Paragraph-length responses
  • 4096 — Detailed explanations
  • 8192 — Long documents, code files

Use threads efficiently

Threads carry conversation history. Longer threads consume more neurons per request. Use context_mode on the request to control how much history is included. See Thread Context Modes.

Choose the right model

For simple queries (classification, short answers), gpt-4o or gpt-4o-mini consume fewer resources than gpt-4o.

Usage Monitoring

Track your consumption in real time:

  • Freddy → Organization → Usage
  • Analytics API — Programmatic access to usage data by time range, model, and assistant

See Usage Limits Guide for setting spending limits and alerts.

Free Tier

New organizations receive a trial allocation of synapses and neurons. Check your Freddy dashboard for your current trial balance.