Salesforce Einstein Requests Pricing Update – October 2025
Salesforce has announced an update to the pricing model for Einstein Requests, effective October 24, 2025. This update brings more transparency and control over costs, allowing enterprises to predict, budget, and scale AI usage with confidence.
What's New?
The new pricing model calculates usage based on:
- LLM Type: Starter, Basic, Standard, Advanced
- API Call Size Factor: Calculated using total tokens (input + response)
Example Calculations:
| LLM Type | Tokens | Einstein Requests |
|---|---|---|
| Starter | 1,000 | 4 |
| Standard | 3,500 | 20 |
| Advanced | 10,000 | 190 |
Formula:
Einstein Requests per LLM API Call =
1RoundUp((Total input + output tokens) / 2000) * Usage Type Multiplier
Usage Type Multipliers: Starter/Basic = 4, Standard = 10, Advanced = 38
Key Benefits
Predictable Costs
No more guessing—usage depends on clear factors:
- LLM Type – Determines the multiplier.
- API Call Size – Size of your input + AI response in tokens.
Flexibility
Use Salesforce foundational LLMs or your own LLMs, depending on your workflow requirements.
Efficient Scaling
Integrate AI confidently into your workflows with transparent and predictable usage metrics.
Pricing Table (Effective Oct 24, 2025)
| LLM Type | Usage Type Multiplier | API Call Example | Einstein Requests Consumed |
|---|---|---|---|
| Starter | 4 | 1,000 tokens | 4 |
| Basic | 4 | 2,000 tokens | 4 |
| Standard | 10 | 3,500 tokens | 20 |
| Advanced | 38 | 10,000 tokens | 190 |
Note: The number of Einstein Requests may also consume Data Cloud credits, Flex Credits, or Conversations, depending on the Salesforce features used.
How to Calculate Einstein Requests
- Add the size of the prompt request and LLM response in tokens.
- Divide by 2,000 tokens.
- Round up to the next highest integer.
- Multiply by the usage type multiplier.
Understanding Tokens
- Tokens are the smallest processing units in LLMs.
- Roughly, 2,000 tokens ≈ 1,500 words.
- Token counts vary for documents and images depending on text density, formatting, and content type.
Key Takeaways
- Predictable AI Costs: Plan usage and budget accurately.
- Flexible LLM Options: BYO-LLM or Salesforce foundational models.
- Transparent Usage Formula: Calculate Einstein Requests with ease.
Official Salesforce Source:
For a comprehensive understanding of the new pricing model, refer to the official Salesforce documentation:
