Our voice-to-voice model resolves 84% of customer-service tasks, on par with other frontier models.
 | Introducing Fluid | Today we're launching Fluid, our voice-to-voice model built to solve problems. You're on this list, so you're hearing it before anyone else. On τ²-bench, the customer-service benchmark, Fluid resolves 84% of tasks, on par with other frontier models. It also talks the way people talk. It backchannels while you speak (mhm, aha, etc.) and keeps you posted while it checks something. It's the most natural-sounding model we've built to date. | |  | Why build on Fluid | | • | It delegates. When a conversation needs a lookup, Fluid hands the task to an LLM that uses your tools. | | • | It sounds natural out of the box. | | • | It works in every language and with every voice we already support. | | • | Switching is easy: if you're on OpenAI's Live API, it takes a new URL and key. | | • | Every turn and every tool call also lands in a text log. |
| One more thing | We go live on LinkedIn today at 16:00 Berlin and on PH tomorrow. A comment or a repost in the first hour helps more than you'd think! | | | You receive this email because you subscribed to “Product updates” at KugelAudio. Unsubscribe · Account settings KugelAudio GmbH · Rosenthaler Str. 36 · 10178 Berlin · Germany
Register: Amtsgericht Charlottenburg, HRB 277989 B · VAT ID: DE458999009
Managing directors: Kajo Kratzenstein, Viktor Presber |
|
|