# How Parallel Web Systems uses Baseten

**3x**: “3x cost savings when moving from closed-source to open-source models”

As published on [baseten.co](https://www.baseten.co/resources/customers/how-parallel-web-systems-pioneers-programmatic-web-with-high-throughput-agentic-i/), July 2026. Captured by usedby on 2026-10-05.

- Company: [Parallel Web Systems](https://www.usedby.ai/companies/parallel-web-systems.md)
- Tool: [Baseten](https://www.usedby.ai/tools/baseten.md)

## What the story says

Parallel runs the multi-step agentic pipelines behind its web search and research APIs on Baseten. It started on pay-per-token Model APIs, then moved to dedicated deployments of an open-source model and now also trains and deploys fine-tuned models on the platform.

Summary written by usedby from the source page, in English. The figures are those of Baseten and Parallel Web Systems, not ours.

> Baseten’s throughput and latency optimizations delivered 3x savings versus the equivalent closed-source model. Eliminating redundant processing made it economically viable to run the kind of long, multi-turn agent workflows that deliver meaningfully better results than a single-shot query.

- **50%**: “Baseten’s proprietary runtime cut latency by 50% and increased throughput by 3x through KV cache-aware routing and speculative decoding.”
- **~100x**: “Within a few months of their initial deployment, Parallel scaled traffic with Baseten ~100x.”

> Our workloads can be very spiky. We need to be able to manage the peaks, but we also don't want to pay for the valleys. Pay-per-token with Model APIs enabled us to easily make the transition to open-source until our scale merited dedicated deployments optimized for throughput and flexible autoscaling.
>
> Matt Lee, Engineering Lead

## What usedby checked

We compared the story with its live page on 2026-10-05.

- Checked: the figure 3x is printed word for word on the page, near the name of Parallel Web Systems.
- Checked: the passage quoted above is copied word for word from the page, near the name of Parallel Web Systems.
- Checked: the publication date is read from the page’s own metadata, never guessed.
- Checked: Parallel Web Systems uses Baseten. Confirmed line. Latest check across sources: 2026-10-05.
- Not checked: the result itself. We quote it; we did not measure it.

---
Source: https://www.usedby.ai/case-studies/parallel-web-systems-baseten · How we check: https://www.usedby.ai/methodology