# AlphaSense speeds up generative search with Cerebras hardware

> AlphaSense’s new ‘auto’ mode returns research answers in 10‑20 seconds, thanks to faster model calls on Cerebras chips.

Oossa · 2026-09-30 · https://oossa.com/en/alphasense-speeds-up-generative-search-with-cerebras-hardware

Produced and translated with AI assistance. Check the original sources below.

AlphaSense announced that its Generative Search now runs on Cerebras’s high‑speed inference chips. The partnership cuts the time to first token by 88%, from 19.5 seconds to 2.3 seconds, letting the system deliver initial answers in as little as 10‑20 seconds. Faster model calls let AlphaSense keep the multi‑step research workflow—planning, routing, snippet evaluation—while still feeling interactive.

## The facts

- Cerebras reduces p90 time to first token from 19.5 s to 2.3 s (8.5× faster).
- AlphaSense’s ‘auto’ mode can return results in 10‑20 s.

## Why it matters

Analysts can get evidence‑based answers within minutes instead of hours, letting them iterate faster on earnings calls and market research.

## Sources & references

1. [How AlphaSense Uses Fast Inference to Make Agentic Research Interactive September 30, 2026](https://www.cerebras.ai/blog/how-alphasense-uses-fast-inference-to-make-agentic-research-interactive) – Cerebras

Last updated: 2026-09-30