1 Release published by 1 person

• v1.27.0

  published Jul 23, 2025

10 Pull requests merged by 2 people

• feat: Define StringDict type alias for better readability

  #1168 merged Jul 24, 2025

• feat: allow methods in tuning to return headers in sdk_http_response by default

  #1165 merged Jul 24, 2025

• feat: allow methods in models to return headers in sdk_http_response by default.

  #1107 merged Jul 24, 2025

• fix: Retry async when seeing aiohttp ClientConnectorError, ClientOSError, ServerDisconnectedError

  #1159 merged Jul 23, 2025

• chore: Update docs

  #1170 merged Jul 23, 2025

• chore(main): release 1.27.0

  #1139 merged Jul 23, 2025

• feat: return headers for list method in all modules.

  #1151 merged Jul 22, 2025

• docs: Copy improvements and minor changes to instructions.

  #1158 merged Jul 22, 2025

• chore: Enable replay tests for Async Java batches module

  #1152 merged Jul 21, 2025

• chore: move codegen_instructions.md

  #1160 merged Jul 21, 2025

10 Pull requests opened by 6 people

• chore: Set default location to us-central1 if project is set

  #1153 opened Jul 19, 2025

• chore: Set credentials from environment if not provided

  #1154 opened Jul 19, 2025

• docs: add comprehensive LLM-friendly documentation

  #1155 opened Jul 19, 2025

• chore: Renamed internal method for get Operation

  #1161 opened Jul 21, 2025

• Update test_create.py

  #1163 opened Jul 22, 2025

• Update errors.py

  #1167 opened Jul 23, 2025

• fix: Allow empty proj/location and api key when overriding base_url

  #1169 opened Jul 23, 2025

• chore(main): release 1.28.0

  #1171 opened Jul 23, 2025

• feat: Support retry configuration at request level

  #1173 opened Jul 24, 2025

• chore: Retry once when catching DefaultCredentialsError

  #1175 opened Jul 25, 2025

4 Issues closed by 2 people

• Response pydantic model containing field of Any type or Union of types will throw an error by gemini api

  #861 closed Jul 24, 2025

• Setting timeout in genai.Client() does not work

  #911 closed Jul 24, 2025

• artifacts appear in Imagen 4.0 image upscaling when using image_bytes

  #972 closed Jul 24, 2025

• Generating a video using a local image input

  #893 closed Jul 18, 2025

10 Issues opened by 9 people

• Set SafetySetting across all categories

  #1174 opened Jul 24, 2025

• Files API should accept TTL shorter than 48 hours

  #1172 opened Jul 24, 2025

• Live API in response_modalities = AUDIO doesn't want to call tools w/o voice acknowledge

  #1166 opened Jul 23, 2025

• Feature Request: File upload progress indicator

  #1164 opened Jul 22, 2025

• ### 🐞 Bug Report: `json.loads(chunk)` fails on Gemini streaming error message

  #1162 opened Jul 22, 2025

• Unexpected truncation of thought summary with Gemini 2.5 Pro

  #1157 opened Jul 21, 2025

• Receiving tool call information with automatic function calling enabled

  #1156 opened Jul 21, 2025

• Batch mode does not support response_schema in JSONL input file

  #1150 opened Jul 18, 2025

• Add HttpRetryOptions example and defaults to documentation

  #1149 opened Jul 18, 2025

• New response_schema invalidates implicit cache

  #1148 opened Jul 18, 2025

64 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Providing thought signatures back DOES NOT change input tokens count

  #1129 commented on Jul 18, 2025 • 0 new comments

• `$ref` JSONSchema keyword is not supported in function definitions

  #1122 commented on Jul 18, 2025 • 0 new comments

• Attach session id to the AsyncSession object

  #1110 commented on Jul 18, 2025 • 0 new comments

• `const` JSONSchema keyword is not supported in function definitions

  #1093 commented on Jul 18, 2025 • 0 new comments

• Function call args are not streamed in chunks

  #1092 commented on Jul 18, 2025 • 0 new comments

• Allow `allowed_tools` for MCP function calling

  #1089 commented on Jul 18, 2025 • 0 new comments

• Update the README.md to include uv installation instructions

  #1041 commented on Jul 18, 2025 • 0 new comments

• Structured Output Returns None When max_output_tokens Exceeded

  #1039 commented on Jul 18, 2025 • 0 new comments

• ConnectionClosedError: 1011 Internal error encountered during live session with Gemini live 2.5 native

  #985 commented on Jul 18, 2025 • 0 new comments

• Grounded Search with Gemini doesn't properly include `part_index` in `GroundingMetadata.grounding_supports.segment`.

  #961 commented on Jul 18, 2025 • 0 new comments

• Support a larger number of enum values when using structured output

  #950 commented on Jul 18, 2025 • 0 new comments

• Python tool calling fails with `400 INVALID_ARGUMENT` when using a `functools.partial` tool

  #907 commented on Jul 18, 2025 • 0 new comments

• Video Understanding is not working for video uploaded through File API, while the same youtube video link works.

  #864 commented on Jul 18, 2025 • 0 new comments

• Possible to return BOTH structured (application/json) and unstructured (markdown) output?

  #839 commented on Jul 18, 2025 • 0 new comments

• Why veo2 is available but not veo3 from the API?

  #860 commented on Jul 18, 2025 • 0 new comments

• Large PDF documents take too long to process

  #809 commented on Jul 18, 2025 • 0 new comments

• "ø" character triggers 2.5 pro response failures...

  #807 commented on Jul 18, 2025 • 0 new comments

• max audio files in a single request

  #758 commented on Jul 18, 2025 • 0 new comments

• Make websockets an optional dependency

  #765 commented on Jul 18, 2025 • 0 new comments

• Getting INVALID_ARGUMENT for some mime types

  #744 commented on Jul 18, 2025 • 0 new comments

• [Docs] Clarity on Multimodal Embeddings Endpoint

  #703 commented on Jul 18, 2025 • 0 new comments

• User location is not supported for the API use.

  #693 commented on Jul 18, 2025 • 0 new comments

• Incorrect list of models and details when using vertexai=True

  #679 commented on Jul 18, 2025 • 0 new comments

• Error handling with gemini Live

  #668 commented on Jul 18, 2025 • 0 new comments

• Support service account impersonation for Vertex AI

  #627 commented on Jul 18, 2025 • 0 new comments

• Error when response schema definition has anyOf with null

  #625 commented on Jul 18, 2025 • 0 new comments

• Add min_p

  #608 commented on Jul 18, 2025 • 0 new comments

• [Ferature Request] Support Google ModelGarden Models

  #567 commented on Jul 18, 2025 • 0 new comments

• Vertex API does not return input_token_limit, Gemini API does

  #574 commented on Jul 18, 2025 • 0 new comments

• Fix documentation

  #556 commented on Jul 18, 2025 • 0 new comments

• Support fixed-length tuples

  #555 commented on Jul 18, 2025 • 0 new comments

• Cannot send cached content to a batch job.

  #509 commented on Jul 18, 2025 • 0 new comments

• Feature Request: Support for per-content titles in embed_content API

  #507 commented on Jul 18, 2025 • 0 new comments

• Add tool_config Feature to LiveConnectConfig

  #468 commented on Jul 18, 2025 • 0 new comments

• Count tokens with `CountTokensConfig` passed in throws 400 INVALID_ARGUMENT error

  #432 commented on Jul 18, 2025 • 0 new comments

• google.genai only support batch embedding which leads to rate limit errors

  #427 commented on Jul 18, 2025 • 0 new comments

• Support running countTokens locally

  #363 commented on Jul 18, 2025 • 0 new comments

• Http Logs/Traces

  #195 commented on Jul 18, 2025 • 0 new comments

• Midsize PDF file yields Error 400: "The request's total referenced files bytes are too large to be read"

  #308 commented on Jul 18, 2025 • 0 new comments

• Any generation request with `video_metadata` results in a 500 error

  #854 commented on Jul 19, 2025 • 0 new comments

• wrong token count in videos

  #1100 commented on Jul 21, 2025 • 0 new comments

• function calling is not working for gemini-2.5-flash-preview-native-audio-dialog

  #843 commented on Jul 21, 2025 • 0 new comments

• Gemini 2.5 TTS cuts off at a certain point.

  #922 commented on Jul 22, 2025 • 0 new comments

• I am trying to integrate gemini batch endpoint in my wrapper, I use google-genai, I wanted to confirm whether batch mode in gemini supports only GCS or BigQuery as source

  #1136 commented on Jul 22, 2025 • 0 new comments

• `parameters_json_schema` seems not supported by Gemini Live API

  #1147 commented on Jul 22, 2025 • 0 new comments

• APIError and subclasses are not unpicklable

  #1144 commented on Jul 22, 2025 • 0 new comments

• GenerateContentResponse.parsed type should support list[pydantic.BaseModel]

  #882 commented on Jul 22, 2025 • 0 new comments

• Is there anyway to set resolution of image

  #1087 commented on Jul 23, 2025 • 0 new comments

• BlockedReason.OTHER problem in 1.4.0+

  #592 commented on Jul 23, 2025 • 0 new comments

• System Instruction Error

  #425 commented on Jul 23, 2025 • 0 new comments

• Accept JSON schema `examples` annotation

  #1094 commented on Jul 23, 2025 • 0 new comments

• Unable to list models

  #1126 commented on Jul 23, 2025 • 0 new comments

• Batch queries can't be configured.

  #1137 commented on Jul 23, 2025 • 0 new comments

• No support for thinking budget in batch request API.

  #1103 commented on Jul 23, 2025 • 0 new comments

• Thinking models are unreliable when `max_output_tokens` set due to them ignoring the thinking budget

  #782 commented on Jul 23, 2025 • 0 new comments

• Connection Error while using Vertex AI API key?

  #1007 commented on Jul 24, 2025 • 0 new comments

• Support anthropic partner models

  #1134 commented on Jul 24, 2025 • 0 new comments

• Getting MALFORMED_FUNCTION_CALL when tool call contains large payload

  #430 commented on Jul 24, 2025 • 0 new comments

• MALFORMED_FUNCTION_CALL finish reason happens too frequently with vertex AI

  #1120 commented on Jul 24, 2025 • 0 new comments

• Structured Output + Code Execution/Thinking

  #1046 commented on Jul 25, 2025 • 0 new comments

• generate content can't handle properly special characters.

  #591 commented on Jul 25, 2025 • 0 new comments

• Update tenacity library Support

  #1005 commented on Jul 25, 2025 • 0 new comments

• chore(deps): bump urllib3 from 2.2.3 to 2.5.0

  #995 commented on Jul 23, 2025 • 0 new comments

• feat: Support Imagen image segmentation on Vertex

  #1141 commented on Jul 22, 2025 • 0 new comments