Amazon Bedrock FAQs changed AWS, Feb 1, 2026

1. -Removed: Which FMs are available in Amazon Bedrock?
2. -Removed: Why should I use Amazon Bedrock?
3. -Removed: How can I get started with Amazon Bedrock?
4. -Removed: What are the most common use cases for Amazon Bedrock?
5. -Removed: In which AWS Regions is Amazon Bedrock available?
6. -Removed: How do I customize a model on Amazon Bedrock?
7. -Removed: Can I train a model and deploy it on Amazon Bedrock?
8. -Removed: What is latency-optimized inference in Amazon Bedrock?
9. -Removed: How do we get started with latency-optimized inference in Amazon Bedrock?
10. -Removed: What key capabilities does AgentCore provide?
11. -Removed: Which agent frameworks does AgentCore support?
12. -Removed: I am using Amazon Bedrock Agents today. Should I switch to AgentCore?
13. -Removed: Is the content processed by Amazon Bedrock moved outside the AWS Region where I am using Amazon Bedrock?
14. -Removed: Are user inputs and model outputs made available to third-party model providers?
15. -Removed: What security and compliance standards does Amazon Bedrock support?
16. -Removed: Will AWS and third-party model providers use customer inputs to or outputs from Amazon Bedrock to train Amazon Nova, Amazon Titan or any third-party models?
17. -Removed: What SDKs are supported for Amazon Bedrock?
18. -Removed: How much does Amazon Bedrock cost?
19. -Removed: What support is provided for Amazon Bedrock?
20. -Removed: How can I track the input and output tokens?
21. -Removed: Why do I see a billing entry for AWS Marketplace for my usage of AWS Bedrock?
22. -Removed: How can I securely use my data to customize FMs available through Amazon Bedrock?
23. -Removed: How does Amazon Bedrock ensure my data used in fine-tuning remains private and confidential?
24. -Removed: Does Amazon Bedrock support continued pretraining?
25. -Removed: Why should I use continued pretraining in Amazon Bedrock?
26. -Removed: How does the continued pretraining feature relate to other AWS services?
27. -Removed: How do I use continued pre-training?
28. +Added: We recommend you use Amazon Bedrock for model fine-tuning when you:
29. +Added: Are a generative AI application builder who wants a managed API-driven approach, fewer hyperparameters, and an abstraction of the complexity associated with model training.
30. +Added: Require minimal infrastructure overhead, have little to no existing ML infrastructure investments to consider, and are looking to deploy quickly in a serverless manner.
31. +Added: We recommend you use Amazon SageMaker AI for model fine-tuning when you:
32. +Added: Are a data scientist, ML engineer, or AI model developer who wants access to advanced customization techniques such as knowledge distillation, supervised fine tuning, or direct preference optimization, for both full weights and parameter efficient fine-tuning. SageMaker AI also provides the ability to customize your training recipe and model architecture.
33. +Added: Have established ML workflows and infrastructure investments aimed at having greater control over infrastructure and cost.
34. +Added: Want greater flexibility to bring your own libraries and frameworks to optimize training workflows for better accuracy and performance.
35. -Removed: Where can I learn more about the data processed to develop and train Amazon Titan FMs?
36. -Removed: Which data sources can I connect to Amazon Bedrock Knowledge Bases?
37. -Removed: How does Amazon Bedrock Knowledge Base retrieve data from structured data sources?
38. -Removed: Does Amazon Bedrock Knowledge Bases support multi-turn conversations?
39. -Removed: Does Amazon Bedrock Knowledge Bases provide source attribution for retrieved information?
40. -Removed: What multi-modal capabilities does Amazon Bedrock Knowledge Bases offer?
41. -Removed: What multi-modal data formats does Amazon Bedrock Knowledge Bases support?
42. -Removed: What are the different parsing options available in Amazon Bedrock Knowledge Bases?
43. -Removed: How does Amazon Bedrock Knowledge Bases ensure data security and manage workflow complexities?
44. -Removed: What is Model Evaluation on Amazon Bedrock?
45. -Removed: Against what metrics can I evaluate FMs?
46. -Removed: What is the difference between human-based and automatic evaluations?
47. +Added: Amazon Bedrock Guardrails provides configurable safeguards to help safely build generative AI applications at scale. With a consistent and standard approach used across a wide range of foundation models (FMs) including FMs supported in Amazon Bedrock, fine-tuned models, and models hosted outside of Amazon Bedrock, Guardrails delivers industry-leading safety protections for your generative AI applications.
48. +Added: Amazon Bedrock Guardrails offers six safeguards to help you build safe, generative AI applications. Below are the safeguards offered by Bedrock Guardrails.
49. +Added: Multi modal content filters - Configure thresholds to help detect and filter harmful text and/or image content across multiple categories including hate, insults, sexual, violence, misconduct, and prompt attacks.
50. -Removed: Amazon Bedrock Guardrails helps you implement safeguards for your generative AI applications based on your use cases and responsible AI policies. You can create multiple guardrails tailored to different use cases and apply them across multiple foundation models (FMs), providing a consistent user experience and standardizing safety and privacy controls across your generative AI applications.
51. -Removed: What are the safeguards available in Amazon Bedrock Guardrails?
52. -Removed: Guardrails help you to define a set of six policies to help safeguard your generative AI applications. You can configure the following policies in Amazon Bedrock Guardrails:
53. -Removed: Multi modal content filters - Configure thresholds to detect and filter harmful text and/or image content across categories including hate, insults, sexual, violence, misconduct, and prompt attacks.
54. +Added: Bedrock Guardrails supports both text and image content to enable customers to build safe generative AI applications at scale.
55. -Removed: What modalities are supported with Bedrock Guardrails?
56. -Removed: Bedrock Guardrails supports both text and image content to enable customers build secure generative AI applications at scale.
57. -Removed: Can I use Guardrails with all available FMs and tools on Amazon Bedrock?
58. +Added: Bedrock Guardrails provides safeguard tiers for content filters and denied topics with distinct performance characteristics and expanded language support for different application requirements and use cases. There are two tiers with Bedrock Guardrails: Standard tiers that provide robust performance with comprehensive language support. This tier requires opting into cross-region inference. Classic tier offers established functionality and limited language support of 3 languages. See the user guide for more details.
59. +Added: Languages support by Amazon Bedrock Guardrails depends on the filter and the tier used. See the user guide for details on the languages supported for each filter and tier. Any language that is not supported in either classic or standard tier will lead to ineffective results.
60. +Added: Amazon Bedrock Guardrails detects and protects against the following types of prompt attacks.

45 more changed paragraphs are on the page itself.