A simple, transparent pricing model

Gumloop customers have started to scale their usage beyond even our own expectations. Cost clarity is becoming more and more critical as we evolve into core infrastructure for our customers.

We're introducing a new pricing model that is simple, scalable, and transparent.

Here's what is changing:

• We pass through tokens at cost
• We pass through compute at cost
• We charge a base 8% orchestration fee on top of these costs

Why are we doing this?

We’ve noticed a pattern emerging in how customers approach planning their long term AI strategy. There is a clear divide between the tools they consider to be 3rd-party vendors and their "internal infrastructure."

Products like Snowflake, AWS, Vercel, and Databricks are external vendors but have become so core to a company’s offering, that they’re effectively internal infrastructure.

Our customers already deploy Gumloop in their own clouds, integrate hundreds of internal data sources and roll it out wall-to-wall. The only thing preventing us from crossing this internal to external chasm is our pricing.

The future of AI software pricing

Our prediction is that most platforms will transition to this sort of passthrough pricing. The ones that don’t will generate much more revenue in the short term, but will eventually die out.

"Do we build this in-house or buy something?" will become more common with every step change in model performance. The only way to last as a partner to scaling customers is to make it clear how much it would cost to run internally and provide value well beyond what you charge. 

What will change for users

For every task automated by an agent, you will see a much clearer breakdown of what the cost was composed of. The compute that was used to run the agent, the inference and the tools called. On top of this, you’ll clearly see Gumloop's orchestration fee for every task.  

For exact details on our new pricing, visit our pricing page.