Akamai Inference launches in beta with 12 open-weight models

On October 27, 2026, Akamai introduced Akamai Inference, a managed service for open-weight and custom models on Akamai Cloud, alongside Akamai Agent Sandbox for running agent code in isolation. Both are in beta.

What changed

• Akamai Inference serves open-weight and custom models through OpenAI-compatible APIs.
• Two ways to buy: serverless pay-per-use, or dedicated deployments. The beta includes the serverless option.
• The launch catalog has 12 open-weight models from 8 billion to 2.8 trillion parameters, including NVIDIA Nemotron 3 Ultra, OpenAI gpt-oss-120b and Google Gemma 4, plus embedding models.
• Akamai added NVIDIA Blackwell Ultra GPU clusters to its dedicated fleet (B300 clusters through Akamai sales) and says it is preparing for NVIDIA Vera Rubin in 2027.
• Akamai Agent Sandbox runs agent-written code in WebAssembly isolation, as an open source part of the Spin framework.

The announcement does not list per-token prices, rate limits or the full model list.

Hosted prices for the named models

RouterPlus catalog, November 1, 2026, per million tokens. These are prices from RouterPlus providers, not Akamai's.

Compare every model and provider on the RouterPlus models page.

What to do

• If you want to try Akamai, ask for beta access and get per-token prices in writing before you move traffic.
• Run your own evals on any new host before switching. The same weights can behave differently across providers.
• To call these models today through one key:

curl https://api.routerplus.com/v1/chat/completions \
  -H "Authorization: Bearer $ROUTERPLUS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "openai/gpt-oss-120b", "messages": [{"role": "user", "content": "Hello"}]}'

Sign up at app.routerplus.com/login for an API key. Marketplace calls use prepaid credits at the listed rates.

Sources

1. Akamai: Akamai Launches AI Inference Service and Agent Execution Environment
2. RouterPlus model catalog