LLM hosting prices: what changed in September 2026

No price cuts and no rises. Three models newly sold per token.

0 price cuts · 0 price rises · 3 models newly sold per token · 0 withdrawn · 0 new GPU instances · 0 new regions

Models newly sold per token

AWS: Grok 4.7 from $1.00 in / $3.00 out per 1M tokens
AWS: Claude Sonnet 5.5 (Amazon Bedrock Edition) from $2.00 in / $10.00 out per 1M tokens
AWS: OpenAI GPT-6 Astra (Amazon Bedrock Edition) from $10.00 in / $50.00 out per 1M tokens

AWS: OpenAI GPT-6 Astra now supports UltraFast mode on Amazon Bedrock
Azure: Retirement: NVv3-series Azure Virtual Machines
Azure: Retirement: NVv4-series Azure Virtual Machines
AWS: Amazon Bedrock expands Claude models in-region support in the UK (London)
AWS: Amazon Bedrock Managed Agents, powered by OpenAI, is now available in preview
AWS: OpenAI GPT-6.1 Sol is now generally available on Amazon Bedrock
AWS: Amazon Bedrock expands Claude model availability to India, South Korea, and Singapore
AWS: Grok 4.7 is now available on Amazon Bedrock
OCI: Use xAI Grok 4.7 in OCI Generative AI
OCI: Import Z.ai GLM-5.3-Flash into OCI Generative AI
AWS: Amazon SageMaker HyperPod Inference Gateway for scalable LLM inference
OCI: Import Google Gemma 4 26B A4B IT into OCI Generative AI
OCI: Import DeepSeek V4.1 Flash into OCI Generative AI
OCI: Find models by region with model discovery in OCI Generative AI