Cognition AI Deploys NVIDIA Vera Rubin, Claims 4.8x Inference Speedup Over Blackwell
Sign in✓ Signed in
Cognition AI has become the first production customer to run NVIDIA's next-generation Vera Rubin inference chip on CoreWeave's cloud platform, demonstrating a substantial performance leap for autonomous AI agents. The company reported a 4.8-fold increase in token throughput for its SWE-2 autonomous software engineer model on Vera Rubin compared to NVIDIA's current-generation GB200 processor, measured at equivalent latency levels.
Continue reading
The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in✓ Signed in — this article isn’t included in your current plan.
Unlocking the full article...