Akamai Inference
Akamai's managed inference service with OpenAI-compatible APIs, 12 launch models from 8B to 2.8T parameters, and Blackwell GPUs.
Overview
Akamai Inference, announced October 27, 2026 and in beta, gives managed access to open-weight and custom models through OpenAI-compatible APIs, as serverless pay-per-use inference or dedicated deployments. The launch catalog has 12 open-weight models from 8B to 2.8T parameters, including NVIDIA Nemotron 3 Ultra, gpt-oss-120b and Gemma 4, plus embedding models for RAG. Akamai also offers Blackwell Ultra clusters with eight-GPU nodes (288 GB HBM3e per GPU) for tuning and self-hosting, NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs for low-latency inference, and says it will prepare data centers for NVIDIA Vera Rubin in 2027.
This is a sourced directory entry. Features and capabilities are reported in the linked sources; this entry does not document independent Aymorphic testing.
Key features
- OpenAI-compatible APIs with serverless pay-per-use or dedicated deployments
- 12 open-weight launch models, 8B to 2.8T parameters, plus embedding models
- Blackwell Ultra B300 nodes: eight GPUs, 288 GB HBM3e each (2.3 TB per node)
- RTX PRO 6000 Blackwell Server Edition for low-latency inference
- Vera Rubin data-center preparation planned for 2027
Use cases
- Serving open-weight models behind an OpenAI-compatible endpoint without owning GPUs
- Self-hosting or tuning large models on dedicated B300 clusters
- Edge-adjacent inference for latency-sensitive applications
Pricing
Unknown; the press release does not state pricing. Beta, with B300 cluster availability via Akamai Sales.
Pricing reflects the source record and may change. Confirm current terms with the provider.
Dates and sources
- Released
- 27 Oct 2026
- Record updated
- 2 Nov 2026
- Record last checked
- 2 Nov 2026