AWS general documentation change
Summary
Updated service quotas for Advanced Prompt Optimization jobs, guardrails, model inference tokens, and Knowledge Bases. Added new quotas for Anthropic Claude Opus 4.7 and Writer AI Palmyra models. Removed outdated model quotas.
Security assessment
Changes involve routine updates to service quotas and limits, including guardrail processing capacities which are security features. However, there is no evidence of security vulnerabilities being addressed - changes are operational adjustments to throughput limits and region availability. Guardrail updates increase processing capacity for content/sensitive information filters but don't introduce new security functionality.
Diff
diff --git a/general/latest/gr/bedrock.md b/general/latest/gr/bedrock.md index 2ed66a212..3ff171f4b 100644 --- a//general/latest/gr/bedrock.md +++ b//general/latest/gr/bedrock.md @@ -161,2 +161,2 @@ Name | Default | Adjustable | Description -(Advanced Prompt Optimization) Active jobs per account | Each supported Region: 20 | No | The maximum number of active Advanced Prompt Optimization (APO) jobs per account. -(Advanced Prompt Optimization) Inactive jobs per account | Each supported Region: 5,000 | No | The maximum number of inactive Advanced Prompt Optimization (APO) jobs per account. +(Advanced Prompt Optimization) Active jobs per account | Each supported Region: 20 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-B46DD052) | The maximum number of active Advanced Prompt Optimization (APO) jobs per account. +(Advanced Prompt Optimization) Inactive jobs per account | Each supported Region: 5,000 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-0B66D421) | The maximum number of inactive Advanced Prompt Optimization (APO) jobs per account. @@ -290 +290 @@ Name | Default | Adjustable | Description -(Guardrails) On-demand ApplyGuardrail Content filter policy text units per second (standard) | us-east-1: 200 us-west-2: 200 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-6B3D3DE4) | The maximum number of text units that can be processed for Content filter policies per second. This applies to the standard content policy model. +(Guardrails) On-demand ApplyGuardrail Content filter policy text units per second (standard) | us-east-1: 200 us-east-2: 200 us-west-1: 200 us-west-2: 200 ap-northeast-1: 50 ap-northeast-2: 100 ap-south-1: 50 ap-southeast-1: 100 eu-central-1: 50 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-6B3D3DE4) | The maximum number of text units that can be processed for Content filter policies per second. This applies to the standard content policy model. @@ -293,2 +293,2 @@ Name | Default | Adjustable | Description -(Guardrails) On-demand ApplyGuardrail Sensitive information filter policy text units per second | us-east-1: 200 us-west-2: 200 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-CFCAAB0E) | The maximum number of text units that can be processed for Sensitive information filter policies per second -(Guardrails) On-demand ApplyGuardrail Word filter policy text units per second | us-east-1: 200 us-west-2: 200 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-9F4DB459) | The maximum number of text units that can be processed for Word filter policies per second +(Guardrails) On-demand ApplyGuardrail Sensitive information filter policy text units per second | us-east-1: 500 us-east-2: 100 us-west-1: 50 us-west-2: 500 ap-northeast-1: 500 ap-northeast-2: 100 ap-south-1: 200 ap-southeast-1: 100 eu-central-1: 500 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-CFCAAB0E) | The maximum number of text units that can be processed for Sensitive information filter policies per second +(Guardrails) On-demand ApplyGuardrail Word filter policy text units per second | us-east-1: 500 us-east-2: 500 us-west-1: 500 us-west-2: 500 ap-northeast-1: 500 ap-northeast-2: 500 ap-south-1: 500 ap-southeast-1: 500 eu-central-1: 500 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-9F4DB459) | The maximum number of text units that can be processed for Word filter policies per second @@ -296 +296 @@ Name | Default | Adjustable | Description -(Guardrails) On-demand ApplyGuardrail requests per second | us-east-1: 50 us-west-2: 50 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-9072D6F0) | The maximum number of ApplyGuardrail API calls allowed per second +(Guardrails) On-demand ApplyGuardrail requests per second | us-east-1: 100 us-east-2: 100 us-west-1: 100 us-west-2: 100 ap-northeast-1: 100 ap-northeast-2: 100 ap-south-1: 100 ap-southeast-1: 100 eu-central-1: 100 Each of the other supported Regions: 25 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-9072D6F0) | The maximum number of ApplyGuardrail API calls allowed per second @@ -323,0 +324 @@ Name | Default | Adjustable | Description +(Knowledge Bases) Ingestion job file size with text content | Each supported Region: 50 | No | The maximum size (in MB) of a file with text content (such as .txt, .pdf, or .docx files) in an ingestion job. @@ -623 +624 @@ Cross-region model inference tokens per minute for Anthropic Claude Opus 4.6 V1 -Cross-region model inference tokens per minute for Anthropic Claude Opus 4.7 | Each supported Region: 15,000,000 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-5DB28B7B) | The maximum number of cross-region tokens that you can submit for model inference in one minute for Anthropic Claude Opus 4.7. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. +Cross-region model inference tokens per minute for Anthropic Claude Opus 4.7 | Each supported Region: 30,000,000 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-5DB28B7B) | The maximum number of cross-region tokens that you can submit for model inference in one minute for Anthropic Claude Opus 4.7. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. @@ -674,0 +676 @@ Global cross-region model inference tokens per day for Anthropic Claude Opus 4.6 +Global cross-region model inference tokens per day for Anthropic Claude Opus 4.7 | Each supported Region: 43,200,000,000 | No | The maximum number of global cross-region tokens that you can submit for model inference in one day for Anthropic Claude Opus 4.7. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. @@ -688 +690 @@ Global cross-region model inference tokens per minute for Anthropic Claude Opus -Global cross-region model inference tokens per minute for Anthropic Claude Opus 4.7 | Each supported Region: 15,000,000 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-34152C1D) | The maximum number of global cross-region tokens that you can submit for model inference in one minute for Anthropic Claude Opus 4.7. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. +Global cross-region model inference tokens per minute for Anthropic Claude Opus 4.7 | Each supported Region: 30,000,000 | [Yes](https://console.aws.amazon.com/servicequotas/home/services/bedrock/quotas/L-34152C1D) | The maximum number of global cross-region tokens that you can submit for model inference in one minute for Anthropic Claude Opus 4.7. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. @@ -791,0 +794 @@ Model invocation max tokens per day for Anthropic Claude Opus 4.6 V1 (doubled fo +Model invocation max tokens per day for Anthropic Claude Opus 4.7 (doubled for cross-region calls) | Each supported Region: 21,600,000,000 | No | Daily maximum tokens for model inference for Anthropic Claude Opus 4.7. Combines sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. Doubled for cross-region calls; not applicable in case of approved TPM increase. @@ -842,0 +846,2 @@ Model invocation max tokens per day for Voxtral Small 1.0 (doubled for cross-reg +Model invocation max tokens per day for Writer AI Palmyra X4 V1 (doubled for cross-region calls) | Each supported Region: 108,000,000 | No | Daily maximum tokens for model inference for Writer AI Palmyra X4 V1. Combines sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. Doubled for cross-region calls; not applicable in case of approved TPM increase. +Model invocation max tokens per day for Writer AI Palmyra X5 V1 (doubled for cross-region calls) | Each supported Region: 108,000,000 | No | Daily maximum tokens for model inference for Writer AI Palmyra X5 V1. Combines sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream. Doubled for cross-region calls; not applicable in case of approved TPM increase. @@ -936,3 +940,0 @@ On-demand model inference requests per minute for AI21 Labs Jamba 1.5 Mini | Eac -On-demand model inference requests per minute for AI21 Labs Jamba Instruct | Each supported Region: 100 | No | The maximum number of times that you can call model inference in one minute for AI21 Labs Jamba Instruct. The quota considers the combined sum of requests for Converse and InvokeModel -On-demand model inference requests per minute for AI21 Labs Jurassic-2 Mid | Each supported Region: 400 | No | The maximum number of times that you can call InvokeModel in one minute for AI21 Labs Jurassic-2 Mid -On-demand model inference requests per minute for AI21 Labs Jurassic-2 Ultra | Each supported Region: 100 | No | The maximum number of times that you can call InvokeModel in one minute for AI21 Labs Jurassic-2 Ultra @@ -951 +952,0 @@ On-demand model inference requests per minute for Amazon Titan Text Express | Ea -On-demand model inference requests per minute for Amazon Titan Text Lite | Each supported Region: 800 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Amazon Titan Text Lite @@ -960 +960,0 @@ On-demand model inference requests per minute for Anthropic Claude 3.7 Sonnet V1 -On-demand model inference requests per minute for Anthropic Claude Instant | us-east-1: 1,000 us-west-2: 1,000 Each of the other supported Regions: 400 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Anthropic Claude Instant @@ -963,3 +962,0 @@ On-demand model inference requests per minute for Anthropic Claude Sonnet 4.6 | -On-demand model inference requests per minute for Anthropic Claude V2 | us-east-1: 500 us-west-2: 500 Each of the other supported Regions: 100 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Anthropic Claude V2 -On-demand model inference requests per minute for Cohere Command | Each supported Region: 400 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Cohere Command. -On-demand model inference requests per minute for Cohere Command Light | Each supported Region: 800 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Cohere Command Light. @@ -981,4 +977,0 @@ On-demand model inference requests per minute for Magistral Small 1.2 | Each sup -On-demand model inference requests per minute for Meta Llama 2 13B | Each supported Region: 800 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 13B. -On-demand model inference requests per minute for Meta Llama 2 70B | Each supported Region: 400 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 70B. -On-demand model inference requests per minute for Meta Llama 2 Chat 13B | Each supported Region: 800 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 Chat 13B. -On-demand model inference requests per minute for Meta Llama 2 Chat 70B | Each supported Region: 400 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 Chat 70B. @@ -1025,2 +1017,0 @@ On-demand model inference requests per minute for Stability.ai Stable Diffusion -On-demand model inference requests per minute for Stability.ai Stable Diffusion XL 0.8 | Each supported Region: 60 | No | The maximum number of times that you can call InvokeModel in one minute for Stability.ai Stable Diffusion XL 0.8 -On-demand model inference requests per minute for Stability.ai Stable Diffusion XL 1.0 | Each supported Region: 60 | No | The maximum number of times that you can call InvokeModel in one minute for Stability.ai Stable Diffusion XL 1.0 @@ -1053,3 +1043,0 @@ On-demand model inference tokens per minute for AI21 Labs Jamba 1.5 Mini | Each -On-demand model inference tokens per minute for AI21 Labs Jamba Instruct | Each supported Region: 300,000 | No | The maximum number of tokens that you can submit for model inference in one minute for AI21 Labs Jamba Instruct. The quota considers the combined sum of tokens for Converse and InvokeModel -On-demand model inference tokens per minute for AI21 Labs Jurassic-2 Mid | Each supported Region: 300,000 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for AI21 Labs Jurassic-2 Mid. -On-demand model inference tokens per minute for AI21 Labs Jurassic-2 Ultra | Each supported Region: 300,000 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for AI21 Labs Jurassic-2 Ultra. @@ -1059 +1046,0 @@ On-demand model inference tokens per minute for Amazon Nova Pro | Each supported -On-demand model inference tokens per minute for Amazon Titan Image Generator G1 | Each supported Region: 2,000 | No | The maximum number of tokens that you can provide through InvokeModel in one minute for Amazon Titan Image Generator G1. @@ -1065 +1051,0 @@ On-demand model inference tokens per minute for Amazon Titan Text Express | Each -On-demand model inference tokens per minute for Amazon Titan Text Lite | Each supported Region: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Amazon Titan Text Lite. @@ -1074 +1059,0 @@ On-demand model inference tokens per minute for Anthropic Claude 3.7 Sonnet V1 | -On-demand model inference tokens per minute for Anthropic Claude Instant | us-east-1: 1,000,000 us-west-2: 1,000,000 Each of the other supported Regions: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Anthropic Claude Instant. @@ -1077,3 +1061,0 @@ On-demand model inference tokens per minute for Anthropic Claude Sonnet 4.6 | Ea -On-demand model inference tokens per minute for Anthropic Claude V2 | us-east-1: 500,000 us-west-2: 500,000 Each of the other supported Regions: 200,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Anthropic Claude V2. -On-demand model inference tokens per minute for Cohere Command | Each supported Region: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Cohere Command. -On-demand model inference tokens per minute for Cohere Command Light | Each supported Region: 300,000 | No | The maximum number of times that you can call model inference in one minute. The quota considers the combined sum of Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Cohere Command Light. @@ -1094,4 +1075,0 @@ On-demand model inference tokens per minute for Magistral Small 1.2 | Each suppo -On-demand model inference tokens per minute for Meta Llama 2 13B | Each supported Region: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 13B. -On-demand model inference tokens per minute for Meta Llama 2 70B | Each supported Region: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 70B. -On-demand model inference tokens per minute for Meta Llama 2 Chat 13B | Each supported Region: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 Chat 13B. -On-demand model inference tokens per minute for Meta Llama 2 Chat 70B | Each supported Region: 300,000 | No | The maximum number of on-demand tokens that you can submit for model inference in one minute. The quota considers the combined sum of input and output tokens across all requests to Converse, ConverseStream, InvokeModel and InvokeModelWithResponseStream for Meta Llama 2 Chat 70B.