Commit graph

109 commits

Author SHA1 Message Date
Sameer Kankute
cdeefe85ea Add nova sonic tests 2026-02-02 12:18:43 +05:30
shin-bot-litellm
14a5706131
litellm_fix(test): fix Bedrock tool search header test regression (#20135) 2026-01-31 00:44:47 -08:00
Jay Prajapati
4717f742eb
fix: filter unsupported beta headers for Bedrock Invoke API (#19877)
- Add whitelist-based filtering for anthropic_beta headers
- Only allow Bedrock-supported beta flags (computer-use, tool-search, etc.)
- Filter out unsupported flags like mcp-servers, structured-outputs
- Remove output_format parameter from Bedrock Invoke requests
- Force tool-based structured outputs when response_format is used

Fixes #16726
2026-01-27 17:47:27 -08:00
michelligabriele
7d5439adda
fix(bedrock): support tool search header translation for Sonnet 4.5 (#19871)
Extend advanced-tool-use header translation to include Claude Sonnet 4.5
in addition to Opus 4.5 on Bedrock Invoke API.

When Claude Code sends the advanced-tool-use-2025-11-20 header, it now
gets correctly translated to Bedrock-specific headers for both:
- Claude Opus 4.5
- Claude Sonnet 4.5

Headers translated:
- tool-search-tool-2025-10-19
- tool-examples-2025-10-29

Fixes defer_loading validation error on Bedrock with Sonnet 4.5.

Ref: https://platform.claude.com/docs/en/agents-and-tools/tool-use/tool-search-tool
2026-01-27 12:17:09 -08:00
Sameer Kankute
3f32562587 Translate advanced-tool-use to Bedrock-specific headers for Claude Opus 4.5 2026-01-27 11:20:16 +05:30
Sameer Kankute
12463809bd
Merge pull request #19638 from BerriAI/main
merge main in stagin 1 22 26
2026-01-23 14:54:17 +05:30
Sameer Kankute
24faca9bcf Add support for output formatfor bedrock invoke via v1/messages 2026-01-22 16:36:03 +05:30
João Dinis Ferreira
60840ea292
fix(bedrock): correct streaming choice index for tool calls (#19506)
Bedrock's contentBlockIndex identifies content blocks within a message
(text=0, tool_call=1), not OpenAI's choice index (which varies with n>1).
This caused OpenAI SDK's ChatCompletionAccumulator to fail when tool call
chunks arrived on index 1 while finish_reason arrived on index 0.

Bedrock doesn't support n>1 (no such parameter exists):
https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_InferenceConfiguration.html

OpenAI choice index spec:
https://platform.openai.com/docs/api-reference/chat/streaming
2026-01-21 20:57:14 -08:00
Benedikt Óskarsson
f09cae2107
Merge branch 'main' into fix/bedrock-thinking-tool-call-2 2026-01-19 13:08:17 +00:00
Sameer Kankute
9a27a52424
Merge pull request #18956 from BerriAI/litellm_staging_01_12_2026
Litellm staging 01 12 2026
2026-01-12 18:27:57 +05:30
Sameer Kankute
98f1a0d0c4
Merge pull request #18946 from BerriAI/litellm_add_max_completion_tokens_with_thinking_budget
Fix: respect max_completion_tokens in thinking feat
2026-01-12 18:13:23 +05:30
Sameer Kankute
db32ac217a
Merge pull request #18944 from BerriAI/litellm_fix_bedrock_passthrough_model_id
Fix : model id encoding for bedrock passthrough
2026-01-12 18:08:53 +05:30
Cesar Garcia
c81cd081e9
feat(bedrock): add OpenAI-compatible service_tier parameter translation (#18091)
* feat(bedrock): add OpenAI-compatible service_tier parameter translation

Translates OpenAI's service_tier parameter (string) to Bedrock's
serviceTier format (object with type field).

* docs(bedrock): add OpenAI-compatible service_tier parameter documentation

Document the automatic translation from OpenAI-style service_tier
parameter to Bedrock's native serviceTier format.

* feat(bedrock): add service_tier to response when present

According to OpenAI's API documentation, when service_tier is sent in the
request, it should be returned in the response. This commit implements
this behavior for Bedrock Converse API to maintain compatibility with
OpenAI's API.

Changes:
- Added serviceTier field to ConverseResponseBlock type definition
- Moved ServiceTierBlock definition before ConverseResponseBlock to fix
  type reference order
- Added response transformation to map Bedrock serviceTier (object) to
  OpenAI service_tier (string format)
- Added 4 new tests for response transformation with service_tier

The service_tier is only added to the response when present in Bedrock's
response, maintaining backward compatibility.
2026-01-12 17:28:49 +05:30
Sameer Kankute
508e4da40e Fix: respect max_completion_tokens in thinking feat 2026-01-12 12:44:46 +05:30
Sameer Kankute
dabb459d2b Fix : model id encoding for bedrock passthrough 2026-01-12 10:57:17 +05:30
Sameer Kankute
aa97a34a83 Fix tests/test_litellm/llms/bedrock/test_base_aws_llm.py 2026-01-12 09:05:28 +05:30
Sameer Kankute
a9e57cf272 [Bug]: Add Custom CA certificates to boto3 clients 2026-01-12 09:05:09 +05:30
Ishaan Jaffer
6a9041e67d Revert "aws fix base"
This reverts commit 225f411abc.
2026-01-10 14:08:11 -08:00
Ishaan Jaff
c0cf8bc27d
[Feat] Manus FILES API - Add File upload, get, delete, list (#18904)
* add MANUS get response

* init TwoStepFileUploadRequest

* init TwoStepFileUploadConfig

* add async_create_file to handle 2 step uploads

* init ManusFilesConfig

* add add get_provider_files_config MANUS

* fix validate_environment

* test_manus_files_api_e2e_all_methods

* aws fix base

* init files API MANUS

* test_manus_responses_api_with_file_upload

* mypy lint fixes

* fix BedrockFilesConfig

* manus docs

* docs manus

* mypy lint

* add add fix resposne api utils MANUS
2026-01-10 13:27:54 -08:00
Ishaan Jaffer
3c3ed3bcfb fix resposne api utils 2026-01-10 13:25:31 -08:00
Sameer Kankute
c6101ef977 [Bug]: Add Custom CA certificates to boto3 clients 2026-01-09 12:30:36 +05:30
Sameer Kankute
c78bf8cfcd Add support for model id in bedrock passthrough 2026-01-08 13:03:58 +05:30
Benedikt Óskarsson
81fefc69c9
fix(bedrock): handle thinking with tool calls for Claude 4 models 2026-01-08 00:48:17 +00:00
Sameer Kankute
ed59a9b376 fix background cost tracking tests 2026-01-02 21:12:25 +05:30
Ishaan Jaffer
859474dd44 fix patch 2025-12-20 14:00:42 +05:30
Sameer Kankute
8efd4653ce Fix litellm_mapped_tests_llms 2025-12-19 23:44:14 +05:30
Sameer Kankute
b849f51e58 Add support for stability model in image edit 2025-12-19 23:03:17 +05:30
Sameer Kankute
f83b821f3d fix: Add qwen 2 and qwen 3 in get_bedrock_model_id 2025-12-17 09:22:55 +05:30
Sameer Kankute
6ccfb8b45e Remove ttl fiedl when routing to bedrock 2025-12-16 16:20:19 +05:30
Sameer Kankute
afda51d476
Merge pull request #17873 from BerriAI/litellm_rerank_foraward_headers
Add support for forwarding client headers in /rerank endpoint
2025-12-12 22:25:57 +05:30
Sameer Kankute
d28160b2b2 Add tests for header forwarding 2025-12-12 17:54:17 +05:30
Sameer Kankute
19b48d7bad fix: bedrock header forwarding with cutom api 2025-12-12 16:12:06 +05:30
CyrusTC
5d326386fb
feat(bedrock): add serviceTier support for Converse API (#17810)
Add support for the Bedrock Converse API serviceTier parameter to allow
specifying processing tier (priority, default, or flex).

Changes:
- Add ServiceTierBlock type in litellm/types/llms/bedrock.py
- Add serviceTier to CommonRequestObject
- Add serviceTier to get_config_blocks() in AmazonConverseConfig
- Add comprehensive tests for serviceTier functionality
- Add documentation for serviceTier usage

This allows users to configure service tier via:
- litellm_params in proxy config
- optional_params in SDK calls
2025-12-11 08:16:32 -08:00
Ishaan Jaff
a904067d38
[Feat] New model - add bedrock writer models (#17685)
* add new bedrock models

* test bedrock writer models

* docs bedrock writer palmyra

* add palymra models

* add bedrock writer models

* docs fix
2025-12-08 17:49:06 -08:00
Irfan Sofyana Putra
bffc118170
fix bedrock qwen anthropic beta (#17467) 2025-12-05 15:47:34 -08:00
Sameer Kankute
f2c0029939
Merge pull request #17470 from BerriAI/litellm_batches_bedrock_content
Add support for file content download for bedrock batches
2025-12-04 21:57:04 +05:30
Sameer Kankute
edd392b50d Add support for file content download for bedrock batches 2025-12-04 13:27:53 +05:30
Sameer Kankute
4710e772be Add support for bedrock qwen 2 imported model 2025-12-04 11:08:57 +05:30
flozonn
31cad8e6e6
feat: Add Nova lite 2 reasoning support with reasoningConfig (#17371) 2025-12-02 16:33:07 -08:00
Colin Lin
661bccbc39 fixed flaky test by sorting list 2025-12-01 14:26:14 -05:00
Sameer Kankute
983ba7aa0f Remove not compatible beta header from claude code 2025-12-01 17:22:04 +05:30
Sameer Kankute
7f42b9b987
Merge pull request #17193 from BerriAI/litellm_twelvelabs_int
Added support for twelvelabs pegasus
2025-11-28 22:09:00 +05:30
Sameer Kankute
9d058398df Fix pegasus response and add doc 2025-11-28 21:41:25 +05:30
Andy Forest
8aa4f3d476
fix(bedrock): handle cohere v4 embed response dictionary format (#17220) 2025-11-27 21:50:35 -08:00
Sameer Kankute
247160277e Added support for twelvelabs pegasus 2025-11-27 15:46:04 +05:30
Sameer Kankute
86a9b74fc8 Fix Thinking may not be enabled when tool_choice forces tool use 2025-11-26 13:32:13 +05:30
Sameer Kankute
35bfcac3bc
Add header forwarding in embedding (#16869) 2025-11-24 18:48:10 -08:00
Alan Ponnachan
b92cc2b2f9
fix(bedrock): Ensure consistent chunk IDs in Bedrock streaming responses (#16596)
* ensure consistent chunk IDs in streaming responses

* use native conversationId for consistent stream chunk IDs
2025-11-18 20:37:21 -08:00
Sameer Kankute
b762493ec5
Remove aws params from claude4_5 request (#16315) 2025-11-06 10:10:30 -08:00
Sameer Kankute
887b2e2d98
Add support for empty assitant message in between messages (#15850) 2025-11-01 13:25:21 -07:00