Skip to content

pydantic-ai

v2.12.0 Feature

This release adds 3 notable features for engineering teams evaluating rollout.

βœ“ No known CVEs patched
Read the diff β†’ Tool health β†’ What is this tool? β†’

✓ No known CVEs patched in this version

Topics

agent-framework genai llm pydantic python

Summary

AI summary

Updates πŸ› Bug Fixes, πŸš€ Features, and feat across a mixed release.

Changes in this release

Feature Low

Adds Moonshot AI `kimi-k3` model.

Adds Moonshot AI `kimi-k3` model.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Feature Low

Emits `EnqueuedMessagesEvent` when messages are delivered.

Emits `EnqueuedMessagesEvent` when messages are delivered.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Feature Low

Adds `DeferredToolCallEvent` and `DeferredToolResultEvent` to `AgentStreamEvent`.

Adds `DeferredToolCallEvent` and `DeferredToolResultEvent` to `AgentStreamEvent`.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Bugfix Medium

Fixes `ToolReturnPart` serialization to use field aliases.

Fixes `ToolReturnPart` serialization to use field aliases.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Bugfix Medium

Fixes tool binding for POSITIONAL_OR_KEYWORD params with `*args`.

Fixes tool binding for POSITIONAL_OR_KEYWORD params with `*args`.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Bugfix Medium

Defers streamed output emission when text ends mid-integer-token for Mistral.

Defers streamed output emission when text ends mid-integer-token for Mistral.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Bugfix Medium

Derives capability-contributed toolset `id` from the capability for durable execution.

Derives capability-contributed toolset `id` from the capability for durable execution.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Bugfix Medium

Fixes `GoogleModel` per-Part `media_resolution` forwarding in `vendor_metadata`.

Fixes `GoogleModel` per-Part `media_resolution` forwarding in `vendor_metadata`.

Source: llm_adapter@2026-07-17

Confidence: high

β€”
Full changelog

What's Changed

πŸš€ Features

  • Add Moonshot AI kimi-k3 model by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/6551
  • Emit EnqueuedMessagesEvent when enqueued messages are delivered into the run by @DouweM in https://github.com/pydantic/pydantic-ai/pull/6256
  • feat: add DeferredToolCallEvent and DeferredToolResultEvent to AgentStreamEvent by @YHallouard in https://github.com/pydantic/pydantic-ai/pull/5588

πŸ› Bug Fixes

  • Fix ToolReturnPart serialization to use field aliases so wire output matches return_schema by @pydanty[bot] in https://github.com/pydantic/pydantic-ai/pull/6543
  • Fix tool binding for POSITIONAL_OR_KEYWORD params with *args by @pydanty[bot] in https://github.com/pydantic/pydantic-ai/pull/6541
  • fix(mistral): defer streamed output emission when text ends mid-integer-token by @Mukller in https://github.com/pydantic/pydantic-ai/pull/6519
  • Derive capability-contributed toolset id from the capability for durable execution by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/6408
  • Fix GoogleModel per-Part media_resolution forwarding in vendor_metadata by @pydanty[bot] in https://github.com/pydantic/pydantic-ai/pull/6525

New Contributors

  • @Mukller made their first contribution in https://github.com/pydantic/pydantic-ai/pull/6519
  • @YHallouard made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5588

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v2.11.0...v2.12.0

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track pydantic-ai

Get notified when new releases ship.

Sign up free

About pydantic-ai

AI Agent Framework, the Pydantic way

All releases β†’

Related context

Earlier breaking changes

  • v1.95.0 rename β€˜built-in tools’ to β€˜native tools’, deprecate old fields, register via capabilities=[NativeTool(...)]

Beta — feedback welcome: [email protected]