Skip to content

Releases: pydantic/pydantic-ai

v0.7.5 (2025-08-25)

25 Aug 08:02
f74a95f
Compare
Choose a tag to compare

What's Changed

  • Handle 'STOP' finish_reason in GeminiStreamedResponse by @ArneZsng in #2631
  • Add price() method to ModelResponse by @Kludex in #2584
  • Include thoughts tokens in output_tokens for Google models by @alexmojaki in #2634
  • Add span_id and trace_id to EvaluationReport by @Kludex in #2627
  • Allow proper type on AnthropicProvider when using Bedrock by @akoshel in #2490
  • Use new OpenTelemetry GenAI chat span attribute conventions by @alexmojaki in #2349
  • Ensure content is always set for assistant tool call messages for OpenAI. by @vimota in #2641

New Contributors

Full Changelog: v0.7.4...v0.7.5

v0.7.4 (2025-08-20)

20 Aug 10:07
fde5eea
Compare
Choose a tag to compare

What's Changed

New Contributors

Full Changelog: v0.7.3...v0.7.4

v0.7.3 (2025-08-19)

19 Aug 06:43
10eb5b8
Compare
Choose a tag to compare

What's Changed

  • Deprecate Usage in favour of RequestUsage and RunUsage by @samuelcolvin in #2378
  • Make FallbackModel accept string model names by @vikigenius in #2564
  • Move system_prompt_role from OpenAIModel to OpenAIModelProfile by @Kludex in #2573
  • Add /cp command to CLI to copy last response to clipboard by @07pepa in #2386
  • Pin temporalio to 1.15.0 as plugins API is still experimental by @DouweM in #2582
  • Use _provider.name instead of _system by @Kludex in #2596

New Contributors

Full Changelog: v0.7.2...v0.7.3

v0.7.2 (2025-08-14)

14 Aug 22:55
851df07
Compare
Choose a tag to compare

What's Changed

  • Let message history end on ModelResponse and execute pending tool calls by @DouweM in #2562
  • Ignore leading whitespace when streaming from Qwen or DeepSeek by @DouweM in #2554
  • Stop calling MCP server get_tools ahead of agent run span by @DouweM in #2545
  • Remove anthropic-beta default header set in AnthropicModel by @jerry-reevo in #2544
  • Add OllamaProvider by @DouweM in #2554
  • Add profile and settings to HuggingfaceModel by @DouweM in #2554
  • Ask model to try again if it produced a response without text or tool calls, only thinking by @ethanabrooks in #2556
  • Forward max_uses parameter to Anthropic WebSearchTool by @eballesteros in #2561

New Contributors

Full Changelog: v0.7.1...v0.7.2

v0.7.1 (2025-08-13)

13 Aug 22:20
2293595
Compare
Choose a tag to compare

What's Changed

  • Add new OpenAI GPT-5 models by @medaminezghal in #2503
  • Add support for OpenAI verbosity parameter in Responses API by @ryx2 in #2493
  • Add support for "openai-responses" model inference string by @Kludex in #2528
  • Add UsageLimits.count_tokens_before_request using Gemini count_tokens API by @kauabh in #2137
  • Fix FallbackModel to respect each model's model settings by @jerry-reevo in #2540

New Contributors

Full Changelog: v0.7.0...v0.7.1

v0.7.0 (2025-08-12)

12 Aug 23:14
8f28c34
Compare
Choose a tag to compare

What's Changed

  • Let Agent be run in a Temporal workflow by moving model requests, tool calls, and MCP to Temporal activities by @DouweM in #2225
  • Let toolsets be built dynamically based on run context by @strawgate in #2366
  • Add event_stream_handler to agent and run methods by @DouweM in #2458
  • History processor replaces message history by @AlexEnrique in #2324
  • Add AbstractAgent and WrapperAgent by @DouweM in #2458
  • Add Agent.override(tools=...) by @DouweM in #2458
  • Bump mcp-run-python by @Kludex in #2470
  • Fix error when using GPT-5 with a temperature setting by @DouweM in #2483
  • Fix KeyError when parsing video metadata without audio track in Google models by @jerry-heygen in #2507
  • Make OpenAIResponsesModelSettings.openai_builtin_tools work again by @DouweM in #2520

New Contributors

Full Changelog: v0.6.2...v0.7.0

v0.6.2 (2025-08-07)

07 Aug 15:40
1a22d68
Compare
Choose a tag to compare

What's Changed

New Contributors

Full Changelog: v0.6.1...v0.6.2

v0.6.1 (2025-08-07)

07 Aug 14:59
37a5ddd
Compare
Choose a tag to compare

What's Changed

  • Automatically use OpenAI strict mode for strict-compatible native output types by @DouweM in #2447
  • Make InlineDefsJsonSchemaTransformer public by @DouweM in #2455
  • Send ThinkingParts back to Anthropic used through Bedrock by @DouweM in #2454
  • Support AWS_BEARER_TOKEN_BEDROCK API key env var by @DouweM in #2456
  • Add new Heroku models by @Kludex in #2459

Full Changelog: v0.6.0...v0.6.1

v0.6.0 (2025-08-06)

06 Aug 16:03
13b712f
Compare
Choose a tag to compare

What's Changed

  • Remove older deprecated models and add new model of Anthropic by @medaminezghal in #2435
  • BREAKING CHANGE: Remove next() method from Graph by @Kludex in #2440
  • BREAKING CHANGE: Remove data from FinalResult by @Kludex in #2443
  • BREAKING CHANGE: Remove get_data and validate_structured_result from StreamedRunResult by @Kludex in #2445
  • docs: add griffe_warnings_deprecated by @Kludex in #2444
  • BREAKING CHANGE: Remove format_as_xml module by @Kludex in #2446
  • BREAKING CHANGE: Remove result_type parameter and similar from Agent by @Kludex in #2441
  • Deprecate GoogleGLAProvider and GoogleVertexProvider by @Kludex in #2450
  • BREAKING CHANGE: drop 4 months old deprecation warnings by @Kludex in #2451

Full Changelog: v0.5.1...v0.6.0

v0.5.1 (2025-08-06)

06 Aug 09:33
3ae5e6c
Compare
Choose a tag to compare

What's Changed

  • google: add more information about schema on union by @Kludex in #2426
  • Deprecate GeminiModel in favor of GoogleModel by @Kludex in #2416
  • Use httpx on GoogleProvider by @Kludex in #2438

New Contributors

Full Changelog: v0.5.0...v0.5.1