ainewsblitz.com

Breaking

Kog Unveils GPU Inference Preview Claiming 3,000 Tokens a Second for AI Agents

Kog Unveils GPU Inference Preview Claiming 3,000 Tokens a Second for AI Agents
  • Infra & Chips
  • AI Agents
  • Open Source

French AI infrastructure startup Kog has introduced a technical preview of Kog Inference Engine, claiming that standard data-center GPUs can deliver the low-latency decoding needed for sequential AI-agent workflows. The company reports 3,000 output tokens per second per request on eight AMD Instinct MI300X GPUs and 2,100 on eight NVIDIA H200 GPUs.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,677 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year