4h ago

Cerebras places Kimi K2.6, a trillion-parameter model, into enterprise trials running at roughly 1,000 output tokens per second, the highest speed Artificial Analysis has recorded for any frontier model

——0——

Benchmarks show 981 tokens per second on 10,000 input tokens.

Original post
Lisan al Gaib#980@SCALING01OPCerebrasCerebras@CEREBRAS

Cerebras is now running Kimi K2.6 – a trillion parameter model – in enterprise trials. At ~1,000 tokens/s, this is the fastest frontier model performance ever measured by Artificial Analysis @ArtificialAnlys.

9:44 AM · May 19, 2026 View on X
QUOTE POSTBojan Tunguz#687Bojan Tunguz@TUNGUZ

Holy guacamole!

CerebrasCerebras@cerebras

Cerebras is now running Kimi K2.6 – a trillion parameter model – in enterprise trials. At ~1,000 tokens/s, this is the fastest frontier model performance ever measured by Artificial Analysis @ArtificialAnlys.

4:44 PM · May 19, 2026 · 250.1K Views
5:12 PM · May 19, 2026 · 14.3K Views
QUOTE POSTLisan al Gaib#980Lisan al Gaib@SCALING01

TPUs are insane

Gemini 3.5 Flash is running at ~867 tokens/s almost as fast as Kimi-K2.6 on Cerebras custom chips

CerebrasCerebras@cerebras

Cerebras is now running Kimi K2.6 – a trillion parameter model – in enterprise trials. At ~1,000 tokens/s, this is the fastest frontier model performance ever measured by Artificial Analysis @ArtificialAnlys.

4:44 PM · May 19, 2026 · 250.1K Views
5:38 PM · May 19, 2026 · 4.8K Views