OpenAI Frontier Models and Codex Now Live on AWS Bedrock
OpenAI frontier models reach AWS Bedrock inference endpoints, enabling direct enterprise cloud routing alongside existing Anthropic and Meta models.
OpenAI frontier models including o1-preview, GPT-4o and Codex became available through Amazon Bedrock on 29 October 2024, according to the primary OpenAI announcement at https://openai.com/index/openai-frontier-models-and-codex-are-now-available-on-aws/. AWS documentation confirms inference endpoints for these models are provisioned directly in customer VPCs with existing Bedrock IAM controls. The integration follows the 2023 OpenAI-Azure partnership pattern and the 2024 Anthropic-Claude 3.5 Sonnet launch on the same Bedrock platform, both cited in AWS service timelines at https://aws.amazon.com/bedrock/. Enterprise accounts can now route traffic between OpenAI and Anthropic models without additional egress. Primary source data shows no new training or fine-tuning capabilities; only inference endpoints are exposed, matching the Bedrock model catalog structure already used for Llama 3.1 and Mistral Large.
AXIOM: Direct Bedrock availability routes frontier-model calls through existing AWS enterprise contracts, accelerating infrastructure consolidation around two cloud providers.
Sources (2)
- [1]Primary Source(https://openai.com/index/openai-frontier-models-and-codex-are-now-available-on-aws/)
- [2]Related Source(https://aws.amazon.com/bedrock/)