AWS adds GPT-5.6 to Bedrock, ships one-click Lambda setup and Graviton5 EC2 instances

In its July 20 Weekly Roundup, AWS announced that OpenAI's GPT-5.6 model family — Sol, Terra, and Luna — is now generally available on Amazon Bedrock. Sol is positioned as the flagship for reasoning and cybersecurity, Terra as the balanced everyday option, and Luna as the fast, cost-efficient tier, all running on Bedrock's next-generation inference engine with emphasis on performance, security, and reliability. The move expands frontier-model choice for AWS customers building agentic applications.

Alongside the model news, AWS shipped a one-click Lambda setup prompt to streamline serverless deployment and introduced EC2 C9g and C9gd instances powered by Graviton5, which AWS says deliver 25% better compute performance than Graviton4 with a 5x larger cache. AWS Step Functions also gained an Amazon Bedrock AgentCore-powered agentic reasoning step, embedding AI-agent reasoning directly into workflows.

The Bedrock GPT-5.6 integration and one-click Lambda setup drew developer praise, and Graviton5's gains were welcomed — though some practitioners noted NVIDIA still dominates serious inference workloads and that Graviton's advantage is strongest on general compute rather than large-model serving. Competitively, offering OpenAI's newest models on Bedrock keeps AWS as a neutral model marketplace even as Microsoft evaluates Chinese Kimi K3 for Copilot and Google builds custom silicon.

The pricing angle is notable: Luna's $1/$6 per-million-token pricing has fed 'price war fatigue' on Reddit, where consensus is that Terra is the practical everyday choice while Sol impresses on reasoning. Readers should watch whether Bedrock's GPT-5.6 availability accelerates enterprise migration and how Graviton5 adoption compares against dedicated GPU inference for cost-sensitive agentic workloads.