Key Info

GMI Cloud announced it is a launch partner for EAGLE-4 and has made it available on its NVIDIA B300 fleet from day one, reporting 2.7x faster decoding than plain decoding on Qwen3.8-27B.

Highlights

  • EAGLE-4 is live on GMI Cloud starting today.
  • On the B300 fleet, it delivers 2.7x speedup over plain decoding on Qwen3.8-27B.
  • The upstream EAGLE-4 release reports 961.6 tokens/s on Qwen3.8-27B with TokenSpeed (BF16, 4x NVIDIA B300, TP4, single request, T=0).
  • Full benchmarks from GMI Cloud are promised soon.