@ai-sdk/amazon-bedrock
5.0.145 · 17 Nov 2026Apache-2.06.9M / weekRepositoryHomepagenpmAtom feed
575 stable releases, newest first, by release line. Notes are quoted as the maintainers published them.
- 4.x
4.0.2
23 Dec 2025Only updates dependencies:@ai-sdk/provider-utils 4.0.1@ai-sdk/anthropic 3.0.1
4.0.1
22 Dec 2025Handle
stop_sequence: null
4.0.0
22 Dec 2025- Breaking
Ai SDK 6 beta
provider: Add specificationVersion to ProviderV3
bedrock: Send {} as tool input when streaming tool calls without arguments
Major Changes
- Breaking
Ai SDK 6 beta
Patch Changes
provider: Add specificationVersion to ProviderV3
bedrock: Send {} as tool input when streaming tool calls without arguments
provider/amazon-bedrock: Expose stop_sequence in provider metadata
The Bedrock provider now exposes the specific stop sequence that triggered generation to halt via
providerMetadata.bedrock.stopSequence. This is implemented by:Requesting
/stop_sequenceviaadditionalModelResponseFieldPathsin the API callParsing the value from
additionalModelResponseFields.stop_sequencein both generate and stream responsesExposing it as
stopSequencein the provider metadata (returnsnullwhen no stop sequence was matched)
provider: Shared spec v3
feat(openai); update spec for mcp approval
Fix reasoning with Bedrock when additionalModelRequestFields is used
Fix (provider/amazon-bedrock): deal gracefully with empty tool descriptions
Shorthand names for reranking models
EmbeddingModelV3amazon-bedrock: Clamp temperature to valid 0-1 range with warnings
Support user provided filenames in amazon-bedrock-provider
Tool execution approval
Provider-V3Extended token usage
Added support for reranking models
Add provider version to user-agent header
LanguageModelV3Image editing
Support citations in amazon-bedrock-provider
provider/amazon-bedrock: Normalise headers and body if input is of instance Request
provider/amazon-bedrock,provider/google-vertex-anthropic: Add support for tool calling with structured output
Added support for combining tool calling with structured outputs in both Amazon Bedrock and Google Vertex Anthropic providers. This allows developers to use tools (like weather lookups, web search, etc.) alongside structured JSON output schemas, enabling multi-step agentic workflows with structured final outputs.
Amazon Bedrock Changes:
Removed incorrect warning that prevented using tools with JSON response format
Updated tool choice to use
{ type: 'required' }instead of specific tool selection when using structured outputsAdded
isJsonResponseFromToolparameter to finish reason mappingJSON tool responses are correctly converted to text content and finish reason is mapped from
tool_usetostopAdded comprehensive test coverage for combining tools with structured outputs
Added example files demonstrating the feature
Google Vertex Anthropic Changes:
Inherits support from underlying Anthropic provider implementation
Added test coverage to verify the feature works correctly
Added example files demonstrating the feature
This brings Anthropic provider's structured output capabilities to the Amazon Bedrock and Google Vertex Anthropic providers.
Add Claude Sonnet 4.5 (claude-sonnet-4-5-20250929-v1:0) model support
ImageModelV3amazon-bedrock: Move anthropic_beta to request body
Flexible tool output content support
Expose raw finish reason
provider/bedrock: Support Nova 2 extended reasoning
maxReasoningEffortfieldMoved dependency
@ai-sdk/test-serverto devDependenciesAdd stop sequence support for amazon bedrock provider
tool-specific strict mode
- 3.x
3.0.146
10 Nov 2026Only updates dependencies:@ai-sdk/provider-utils 3.0.46@ai-sdk/anthropic 2.0.115
3.0.145
9 Nov 2026anthropic: Add Claude Fable 5.2 support
Patch Changes
anthropic: Add Claude Fable 5.2 support
add the
claude-fable-5-2model ID to@ai-sdk/anthropicand@ai-sdk/google-vertex,anthropic.claude-fable-5-2andus.anthropic.claude-fable-5-2to@ai-sdk/amazon-bedrock, andanthropic/claude-fable-5.2to@ai-sdk/gatewayrecognize
claude-fable-5-2as a known model with the same capabilities asclaude-fable-5-1: 128k max output tokens, always-on adaptive thinking, no sampling parameters, no forced tool use, and native structured outputs by default
3.0.144
7 Nov 2026Only updates dependencies:@ai-sdk/provider-utils 3.0.45@ai-sdk/anthropic 2.0.113
3.0.143
3 Nov 2026Enable native structured output for supported Claude 5 models on Bedrock. In automatic mode, Bedrock and Anthropic now fall back for schemas that native structured output cannot represent: recursive references, open objects, or more than ten optional properties. Anthropic uses the JSON response tool; Bedrock Converse uses the JSON tool or JSON instructions for models that reject forced tool use. The JSON-tool fallback excludes disabled and unselected application tools. Explicit
outputFormatselection remains unchanged.Default to disabling parallel tool use for named Anthropic tool choices on models that support forced tools. Preserve the complete function tool catalog and prompt cache compatibility. An explicit
providerOptions.anthropic.disableParallelToolUse: falsekeeps parallel calls enabled.amazon-bedrock: Parse JSON string tool inputs and safely serialize BigInt and cycles in tool outputs
Patch Changes
Enable native structured output for supported Claude 5 models on Bedrock. In automatic mode, Bedrock and Anthropic now fall back for schemas that native structured output cannot represent: recursive references, open objects, or more than ten optional properties. Anthropic uses the JSON response tool; Bedrock Converse uses the JSON tool or JSON instructions for models that reject forced tool use. The JSON-tool fallback excludes disabled and unselected application tools. Explicit
outputFormatselection remains unchanged.Default to disabling parallel tool use for named Anthropic tool choices on models that support forced tools. Preserve the complete function tool catalog and prompt cache compatibility. An explicit
providerOptions.anthropic.disableParallelToolUse: falsekeeps parallel calls enabled.amazon-bedrock: Parse JSON string tool inputs and safely serialize BigInt and cycles in tool outputs
3.0.142
30 Oct 2026Only updates dependencies:@ai-sdk/provider-utils 3.0.43@ai-sdk/anthropic 2.0.111
3.0.141
21 Oct 2026Defer validation schema construction until first use to reduce provider import time.
Patch Changes
Defer validation schema construction until first use to reduce provider import time.
3.0.140
7 Oct 2026amazon-bedrock,anthropic-aws: Skip SigV4 signing for non-ASCII header values
3.0.139
7 Oct 2026anthropic: Add Claude Haiku 5.5 support
Patch Changes
anthropic: Add Claude Haiku 5.5 support
add the
claude-haiku-5-5model ID to@ai-sdk/anthropicand@ai-sdk/google-vertex,anthropic.claude-haiku-5-5andus.anthropic.claude-haiku-5-5to@ai-sdk/amazon-bedrock, andanthropic/claude-haiku-5.5to@ai-sdk/gatewayrecognize
claude-haiku-5-5as a known model with a 128k max output token limit and all five effort levels; thinking can be disabled up tohigheffort, andxhighandmaxare lowered tohighwith a warning when thinking is disabledclaude-haiku-5-5does not support thinking token budgets:thinking: { type: 'enabled', budgetTokens }is converted to adaptive thinking with a warning
3.0.138
30 Sep 2026Only updates dependencies:@ai-sdk/provider-utils 3.0.41@ai-sdk/anthropic 2.0.108
3.0.137
28 Sep 2026anthropic: Add Claude Sonnet 5.5 support
Patch Changes
anthropic: Add Claude Sonnet 5.5 support
add the
claude-sonnet-5-5model ID to@ai-sdk/anthropicand@ai-sdk/google-vertex,anthropic.claude-sonnet-5-5andus.anthropic.claude-sonnet-5-5to@ai-sdk/amazon-bedrock, andanthropic/claude-sonnet-5.5to@ai-sdk/gatewayadd the
between_toolsthinking type (thinking: { type: 'between_tools' }), the lowest thinking setting onclaude-sonnet-5-5;xhighandmaxeffort are lowered tohighwith a warning because the API rejects them withbetween_toolsclaude-sonnet-5-5rejects disabled thinking:thinking: { type: 'disabled' }is replaced withbetween_toolsthinking (with a warning), and budget-based thinking is converted to adaptive thinkingclaude-sonnet-5-5rejects forced tool use:requiredand named tool choices fall back toauto, andstructuredOutputMode: 'jsonTool'falls back to native structured outputs, each with a warning
3.0.135 – 3.0.1362 releases that only update dependencies@ai-sdk/anthropic 2.0.106@ai-sdk/provider-utils 3.0.39Show each
3.0.134
23 Sep 2026amazon-bedrock: Avoid forced structured-output tools for incompatible Claude models
Patch Changes
amazon-bedrock: Avoid forced structured-output tools for incompatible Claude models
3.0.133
22 Sep 2026Only updates a dependency:@ai-sdk/anthropic 2.0.103
3.0.132
21 Sep 2026amazon-bedrock: Enable Anthropic features for application inference profile chat models
amazon-bedrock: Preserve API error messages when the error type is omitted
3.0.131
11 Sep 2026amazon-bedrock: Resolve endpoints for non-standard AWS partitions and endpoint overrides
3.0.130
9 Sep 2026Only updates dependencies:@ai-sdk/provider 2.0.4@ai-sdk/provider-utils 3.0.37@ai-sdk/anthropic 2.0.102
3.0.129
4 Sep 2026amazon-bedrock: Return text from non-streaming citation responses
3.0.128
2 Sep 2026Only updates a dependency:@ai-sdk/anthropic 2.0.101
3.0.127
1 Sep 2026Add a structured output mode option and default Claude Sonnet 4.6 and Claude Haiku 4.5 to the JSON tool fallback.
3.0.126
31 Aug 2026Omit assistant messages when only cache points remain after unsigned reasoning is filtered.
Accept citation deltas in Amazon Bedrock streaming responses.
Patch Changes
Omit assistant messages when only cache points remain after unsigned reasoning is filtered.
Accept citation deltas in Amazon Bedrock streaming responses.
3.0.125
27 Aug 2026Only updates a dependency:@ai-sdk/anthropic 2.0.99
3.0.124
26 Aug 2026Forward the Anthropic option for disabling parallel tool use through Amazon Bedrock without sending conflicting tool choice fields.
3.0.120
24 Aug 2026Only updates dependencies:@ai-sdk/provider-utils 3.0.33@ai-sdk/anthropic 2.0.96
3.0.119
20 Aug 2026amazon-bedrock: Surface modeled event-stream exceptions
3.0.118
20 Aug 2026Fix (provider/amazon-bedrock): support
reasoningContent.redactedContentfrom the Converse API and replay it on subsequent turns
3.0.117
19 Aug 20263.0.116
18 Aug 2026bedrock: Wrap invalid tool input in object
3.0.110 – 3.0.1156 releases that only update dependencies@ai-sdk/provider-utils 3.0.32@ai-sdk/anthropic 2.0.95Show each
- 3.0.11511 Aug 2026@ai-sdk/provider-utils 3.0.32@ai-sdk/anthropic 2.0.95
- 3.0.1147 Aug 2026@ai-sdk/anthropic 2.0.94
- 3.0.1133 Aug 2026@ai-sdk/anthropic 2.0.93
- 3.0.11229 Jul 2026@ai-sdk/provider-utils 3.0.31@ai-sdk/anthropic 2.0.92
- 3.0.11124 Jul 2026@ai-sdk/anthropic 2.0.91
- 3.0.11023 Jul 2026@ai-sdk/anthropic 2.0.90
3.0.109
22 Jul 2026Use native structured output on supported Bedrock Anthropic models when thinking is enabled instead of forcing a JSON tool call.
Pass through
s3://image URLs to Amazon Bedrock Converse as S3 image sources instead of downloading them.provider/amazon-bedrock: Omit unsigned reasoning from conversation history
Patch Changes
Use native structured output on supported Bedrock Anthropic models when thinking is enabled instead of forcing a JSON tool call.
Pass through
s3://image URLs to Amazon Bedrock Converse as S3 image sources instead of downloading them.provider/amazon-bedrock: Omit unsigned reasoning from conversation history
Sanitize invalid characters in replayed tool call names before sending conversation history to Amazon Bedrock.
3.0.106 – 3.0.1083 releases that only update dependencies@ai-sdk/anthropic 2.0.88@ai-sdk/provider-utils 3.0.30Show each
3.0.105
30 Jun 2026Feat (provider/anthropic): add
claude-sonnet-5model id
Patch Changes
Feat (provider/anthropic): add
claude-sonnet-5model id
3.0.102 – 3.0.1043 releases that only update dependencies@ai-sdk/provider-utils 3.0.28@ai-sdk/anthropic 2.0.84Show each
3.0.101
9 Jun 2026provider/anthropic: Add support for
claude-fable-5and thefallbacksAPI parameter
Patch Changes
provider/anthropic: Add support for
claude-fable-5and thefallbacksAPI parameter
3.0.100
28 May 2026provider/anthropic: Add support for
claude-opus-4-8
Patch Changes
provider/anthropic: Add support for
claude-opus-4-8
3.0.99
30 Apr 2026Only updates dependencies:@ai-sdk/provider-utils 3.0.25@ai-sdk/anthropic 2.0.79@ai-sdk/provider 2.0.3
3.0.98
29 Apr 2026Trigger release for all packages after provenance setup
Patch Changes
Trigger release for all packages after provenance setup
3.0.97
17 Apr 2026provider/amazon-bedrock: Fix Anthropic reasoning behavior related to Opus 4.7
A whole project to check? Paste its package.json in the dependency check: every deprecated, outdated or vulnerable dependency, with what to upgrade to. Free, no account.