Skip to content

Release history

pydantic-ai releases

AI Agent Framework, the Pydantic way

Back to tool Latest release

All releases

73 shown

No immediate action
v1.105.0 Mixed

On-demand loading + Grok 4.3 + workflow & Google fixes

Open
No immediate action
v1.104.0 New feature

Claude Opus 4.8 support

Open
No immediate action
v1.103.0 Mixed

Features + Bug Fixes

Open
Review required
v1.102.0 Security relevant
RCE / SSRF

SSRF fix + bug fixes

Open
No immediate action
v1.101.0 Mixed

Queue, MCP tasks, XSearch, top_k, bug fixes, dep bump

Open
Review required
v1.100.0 Mixed
RCE / SSRF

SSRF fix + Bedrock JSON

Open
No immediate action
v1.99.0 New feature

gemini-3.5-flash model

Open
No immediate action
v1.98.0 Breaking risk

v2 prep + OpenAI token count + Agent retries

Open
Review required
v1.97.0 Breaking risk
Auth Breaking upgrade

GoogleProvider split and ID rename

Open
No immediate action
v1.96.1 Bug fix

input_fidelity omission + system prompt handling

Open
No immediate action
v1.96.0 Maintenance

v2 deprecations

Open
No immediate action
v1.95.1 Bug fix

Re‑enabled Agent.instrument

Open
Upgrade now
v1.95.0 Breaking risk
Dependencies Breaking upgrade

Native Tool Search

Open
No immediate action
v1.94.0 Breaking risk

Dropped mistralai

Open
v1.93.0 Mixed
Notable features
  • Added `tool_choice` configuration setting
  • Introduced `OutputToolCallEvent` and `OutputToolResultEvent` for handling output tool calls; deprecated corresponding function-tool events
Full changelog

What's Changed

🚀 Features

  • Add tool_choice setting by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/3611
  • Yield OutputToolCallEvent/OutputToolResultEvent for output tool calls; deprecate function-tool events for failing output tool calls by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5320

🐛 Bug Fixes

  • Drain spawned tasks during agent cancellation by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/5341

New Contributors

  • @DhruvGarg111 made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4958

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.92.0...v1.93.0

View release on GitHub
v1.92.0 Mixed
Notable features
  • Added Anthropic task budget support
  • Added runtime `output_retries` override and deprecated `retries`
Full changelog

What's Changed

🚀 Features

  • Add Anthropic task budget support by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/5140
  • Add runtime output_retries override + deprecate retries + internal retry-field rename by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/5075

🐛 Bug Fixes

  • Fix: Clean up streaming responses on cancellation by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/5313
  • Fix attempted exit cancel scope in different task by running MCP session in a dedicated task by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/4514
  • Populate for_run RunContext with run_id, conversation_id, and metadata by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5330
  • fix(evals): guarantee CaseLifecycle.teardown() by @voorhs in https://github.com/pydantic/pydantic-ai/pull/5322

New Contributors

  • @voorhs made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5322

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.91.0...v1.92.0

View release on GitHub
v1.91.0 New feature
Notable features
  • Support for `gpt-image-2` options in OpenAI integration
  • Added support for `deepseek-v4-flash` and `deepseek-v4-pro` models
Full changelog

What's Changed

🚀 Features

  • feat(openai): support gpt-image-2 options by @banteg in https://github.com/pydantic/pydantic-ai/pull/5234
  • feat(deepseek): add support for deepseek-v4-flash anddeepseek-v4-pro by @SuperMarioYL in https://github.com/pydantic/pydantic-ai/pull/5195

🐛 Bug Fixes

  • fix(evals): preserve Unicode in YAML datasets by @Kcstring in https://github.com/pydantic/pydantic-ai/pull/5245
  • Skip empty ModelResponse in OpenAI Chat assistant message mapping by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5300
  • Fix MCP history replay for empty tool arguments by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/5314

New Contributors

  • @Kcstring made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5245
  • @khaldon made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5301
  • @SuperMarioYL made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5195
  • @banteg made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5234

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.90.0...v1.91.0

View release on GitHub
v1.90.0 New feature
⚠ Upgrade required
  • Upgraded web UI dependency to `pydantic/ai-chat-ui` version 1.2.0.
Notable features
  • Added `OpenAIResponsesModelSettings.openai_conversation_id` for OpenAI Conversations API state support.
  • Typed OTel metadata added for code tool call syntax highlighting.
Full changelog

What's Changed

🚀 Features

  • feat: add OpenAI Conversations API state support via OpenAIResponsesModelSettings.openai_conversation_id by @corytomlinson in https://github.com/pydantic/pydantic-ai/pull/5224
  • feat: typed OTel metadata for code tool call syntax highlighting by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5297

📦 Dependencies

  • Upgrade web UI to pydantic/ai-chat-ui 1.2.0 by @johnthagen in https://github.com/pydantic/pydantic-ai/pull/5246

New Contributors

  • @johnthagen made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5246
  • @Acacian made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4888
  • @VeniVeci made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5283

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.89.1...v1.90.0

View release on GitHub
v1.89.1 Bug fix
Notable features
  • Added bundled Library Skills for improved coding‑agent support
Full changelog

What's Changed

🤖 Coding Agent Experience

  • Add bundled Library Skills(https://library-skills.io) for improved coding-agent support by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/5271

🐛 Bug Fixes

  • fix: restore wrap_validation_errors on ToolManager function-tool methods by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5275
  • Defer anyio.Lock creation via cached_property to bind to running loop by @Kludex in https://github.com/pydantic/pydantic-ai/pull/5265

New Contributors

  • @dependabot[bot] made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5268

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.89.0...v1.89.1

View release on GitHub
v1.89.0 New feature
Notable features
  • Added `conversation_id` for cross-run correlation
  • Supported dynamic capabilities via callables in the capabilities list
  • Added `builtin_tools` to `agent.override`
Full changelog

What's Changed

🚀 Features

  • feat: add conversation_id for cross-run correlation by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5251
  • feat(capabilities): support dynamic capabilities via callables in capabilities list by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5252
  • feat: add builtin_tools to agent.override by @mplemay in https://github.com/pydantic/pydantic-ai/pull/5248

🐛 Bug Fixes

  • fix(evals): use non-daemon thread for background evaluators by @alexmojaki in https://github.com/pydantic/pydantic-ai/pull/5247

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.88.0...v1.89.0

View release on GitHub
v1.88.0 Breaking risk
Breaking changes
  • prepare_tools scoped to function tools only; output validate/process hooks and prepare_output_tools introduced
Notable features
  • Cross-provider service_tier model setting supporting Anthropic, Gemini API, Vertex Priority PayGo
  • Fast mode for Anthropic Claude Opus 4.6
  • OpenAI Responses phase support on assistant messages
Full changelog

What's Changed

🚀 Features

  • feat(capabilities)!: output validate/process hooks; scope prepare_tools to function tools, add prepare_output_tools by @DouweM in https://github.com/pydantic/pydantic-ai/pull/4859
  • feat: cross-provider service_tier model setting; Anthropic + Gemini API + Vertex Priority PayGo support by @markmcd in https://github.com/pydantic/pydantic-ai/pull/4926
  • Add anthropic speed, fast mode for opus 4.6 by @bohdanhr in https://github.com/pydantic/pydantic-ai/pull/4300
  • feat: support OpenAI Responses phase on assistant messages by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5229
  • feat(ui): add UIAdapter.sanitize_messages and allowed_file_url_schemes by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5228

🐛 Bug Fixes

  • fix: restore per-block cache_control for anthropic_cache_messages by @Wh1isper in https://github.com/pydantic/pydantic-ai/pull/5227
  • fix: propagate original error when CallToolsNode stream fails by @thejens in https://github.com/pydantic/pydantic-ai/pull/4799
  • fix: propagate Agent(retries=...) to user-provided toolsets by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/4745

New Contributors

  • @markmcd made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4926

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.87.0...v1.88.0

View release on GitHub
v1.87.0 New feature
Notable features
  • HandleDeferredToolCalls capability
  • ProcessEventStream capability
  • GPT-5.5 thinking setting
Full changelog

What's Changed

🚀 Features

  • Add HandleDeferredToolCalls capability and handle_deferred_tool_calls hook by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5142
  • Add ProcessEventStream capability by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5141
  • Handle thinking setting for GPT-5.5 by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/5196

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.86.1...v1.87.0

View release on GitHub
v1.86.1 Bug fix

Minor fixes and improvements.

Full changelog

What's Changed

🐛 Bug Fixes

  • fix(openai): guard against choices=None in streamed chunks (#5165) by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/5170
  • fix(anthropic): send container_id as string on reuse; unwrap broken {id: x} shape by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/5168
  • fix: preserve validation error input on tool-call retries by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5181
  • fix(toolsets): prevent MCPServer.__aexit__ called more times than __aenter__ in DynamicToolset by @anishesg in https://github.com/pydantic/pydantic-ai/pull/5171
  • Warn about dict fields with Anthropic strict mode by @rahulmansharamani14 in https://github.com/pydantic/pydantic-ai/pull/4321
  • Set gen_ai.operation.name=execute_tool on tool-execution spans by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5182
  • fix: avoid stale client references in model and embedding classes by @Deeven-Seru in https://github.com/pydantic/pydantic-ai/pull/4276

New Contributors

  • @anishesg made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5171
  • @rahulmansharamani14 made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4321
  • @Deeven-Seru made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4276

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.86.0...v1.86.1

View release on GitHub
v1.86.0 New feature
Notable features
  • UIAdapter.manage_system_prompt with ReinjectSystemPrompt capability
Full changelog

What's Changed

🚀 Features

  • feat(ui): add UIAdapter.manage_system_prompt + ReinjectSystemPrompt capability by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/4087

🐛 Bug Fixes

  • fix(vercel-ai): allow regenerate requests without messageId by @MukundaKatta in https://github.com/pydantic/pydantic-ai/pull/5156
  • Handle dict reasoning_content from Gemini 2.5 models by @JulieLiu99 in https://github.com/pydantic/pydantic-ai/pull/5155

New Contributors

  • @MukundaKatta made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5156
  • @JulieLiu99 made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5155

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.85.1...v1.86.0

View release on GitHub
v1.85.1 Bug fix

Minor fixes and improvements.

Full changelog

What's Changed

🐛 Bug Fixes

  • fix: exclude validation error input from retry messages by @navalprakhar in https://github.com/pydantic/pydantic-ai/pull/4947
  • fix(vercel-ai): auto-detect deferred tool approval state in dump_messages() by @tijmenhammer in https://github.com/pydantic/pydantic-ai/pull/4831

New Contributors

  • @abhicris made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5149
  • @navalprakhar made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4947
  • @tijmenhammer made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4831

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.85.0...v1.85.1

View release on GitHub
v1.85.0 Maintenance

## What's Changed ### 🚀 Features * Online evaluation via OpenTelemetry events

Full changelog

What's Changed

🚀 Features

  • Online evaluation via OpenTelemetry events by @dmontagu in https://github.com/pydantic/pydantic-ai/pull/5125

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.84.1...v1.85.0

View release on GitHub
v1.84.1 Bug fix

Fixed tool hook behavior for internal output tools and BaseModel arguments.

Full changelog

What's Changed

🐛 Bug Fixes

  • Skip tool hooks for internal output tools by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5128
  • Always pass dict-shaped validated args to hooks for single-BaseModel tools by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5137

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.84.0...v1.84.1

View release on GitHub
v1.84.0 Security relevant
Security fixes
  • Fix exponential-time regex in Google FileSearchTool response parsing
Notable features
  • Claude Opus 4.7 support
  • Stateful compaction mode for OpenAICompaction
  • OllamaModel subclass with corrected capability flags
Full changelog

What's Changed

:lock: Security

  • Fix exponential-time regex in Google FileSearchTool response parsing by @DouweM in #5106. This was first released in yesterday's v1.83.0. The pattern was reachable in principle from Gemini streaming responses but could not be triggered via real Gemini output in testing, so we're treating it as a hardening fix rather than a vulnerability.

🚀 Features

  • Add Claude Opus 4.7 support by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/5118
  • Add stateful compaction mode to OpenAICompaction by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5108
  • Add OllamaModel subclass and correct Ollama capability flags to fix structured output on Ollama Cloud by @Goldokpa in https://github.com/pydantic/pydantic-ai/pull/4160

🐛 Bug Fixes

  • Fix deferred tool search keyword matching by @pandego in https://github.com/pydantic/pydantic-ai/pull/5014
  • Fixing parsing openai_logprobs for streamed responses by @iAmir97 in https://github.com/pydantic/pydantic-ai/pull/5091
  • fix(openai): handle null text in Responses API output by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5124
  • Chain OpenAI Responses retries through concrete openai_previous_response_id seeds by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5126
  • fix: exclude resumed requests from new_messages() even with current run_id by @madanlalit in https://github.com/pydantic/pydantic-ai/pull/4731

New Contributors

  • @iAmir97 made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5091

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.83.0...v1.84.0

View release on GitHub
v1.83.0 New feature
Notable features
  • XSearchTool and FileSearch support for xAI
  • Anthropic automatic prompt caching support
  • Metadata injection per tool call with FastMCPToolset
Full changelog

🤖 Pydantic AI Harness is now live with Code Mode powered by Monty!

What's Changed

🚀 Features

  • feat: Add XSearchTool and FileSearch support for xAI by @colesmcintosh in https://github.com/pydantic/pydantic-ai/pull/4165
  • Allow injecting metadata per tool call with FastMCPToolset by @Boryotto in https://github.com/pydantic/pydantic-ai/pull/4929
  • feat(bedrock): add prompt cache TTL support by @imdigitalashish in https://github.com/pydantic/pydantic-ai/pull/4581
  • Add Anthropic automatic prompt caching support by @DenysMoskalenko in https://github.com/pydantic/pydantic-ai/pull/4840
  • Add 'graceful' end strategy for parallel tool calls by @awkoy in https://github.com/pydantic/pydantic-ai/pull/4333
  • Support Agent(output_type=str | None) for optional output by @Pavanmanikanta98 in https://github.com/pydantic/pydantic-ai/pull/4042

New Contributors

  • @colesmcintosh made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4165
  • @Boryotto made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4929
  • @imdigitalashish made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4581
  • @awkoy made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4333

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.82.0...v1.83.0

View release on GitHub
v1.82.0 New feature
Notable features
  • Pydantic AI Harness with Code Mode powered by Monty
Full changelog

🤖 Pydantic AI Harness is now live with Code Mode powered by Monty!

What's Changed

🐛 Bug Fixes

  • Fix OpenAI compaction chain when openai_previous_response_id='auto' by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5086
  • Fix orphan-task leak in CombinedToolset and CombinedCapability by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5085
  • Lazily compute ToolDefinition.function_signature by @DouweM in https://github.com/pydantic/pydantic-ai/pull/5087

📦 Dependencies

  • chore: bump fastmcp 2.x → 3.2.0 by @dsfaccini in https://github.com/pydantic/pydantic-ai/pull/4982

New Contributors

  • @gvdongen made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5041

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.81.0...v1.82.0

View release on GitHub
v1.81.0 Mixed
Notable features
  • Pydantic AI Harness with Code Mode powered by Monty
Full changelog

🤖 Pydantic AI Harness is now live with Code Mode powered by Monty!

What's Changed

🐛 Bug Fixes

  • Fix Anthropic compaction usage totals by @pandego in https://github.com/pydantic/pydantic-ai/pull/5068
  • Exclude run_id from cache key computation in PrefectAgentInputs by @adtyavrdhn in https://github.com/pydantic/pydantic-ai/pull/3890
  • fix: handle boolean JSON Schema nodes in JsonSchemaTransformer by @Dharit13 in https://github.com/pydantic/pydantic-ai/pull/4989
  • fix: close pending tool calls on stream error by @proever in https://github.com/pydantic/pydantic-ai/pull/4963
  • fix: skip api-version query parameter for Azure AI Foundry serverless endpoints by @Ricardo-M-L in https://github.com/pydantic/pydantic-ai/pull/5071

New Contributors

  • @Dharit13 made their first contribution in https://github.com/pydantic/pydantic-ai/pull/4989
  • @pandego made their first contribution in https://github.com/pydantic/pydantic-ai/pull/5068

Full Changelog: https://github.com/pydantic/pydantic-ai/compare/v1.80.0...v1.81.0

View release on GitHub
v1.80.0 New feature
Notable features
  • CapabilityOrdering with innermost, outermost, wraps, wrapped_by, requires
  • Server-side compaction via OpenAICompaction and AnthropicCompaction
  • Hooks ordering parameter with instance refs
View release on GitHub
v1.79.0 Mixed
Breaking changes
  • HTTP client cache replaced with create_async_http_client context manager
Notable features
  • Full AG-UI 0.1.13 and 0.1.15 support with reasoning and dump_messages
  • AbstractCapability.apply() method
View release on GitHub
v1.78.0 New feature
Notable features
  • Cached token span attributes per OTel specification
  • ToolDefinition.return_schema and function_signature fields
  • SetToolMetadata capability
View release on GitHub
v1.77.0 New feature
Notable features
  • Local WebFetch tool with automatic provider fallback
  • defer_loading parameter for tool search
  • ThreadExecutor capability
View release on GitHub
v1.76.0 Mixed
Breaking changes
  • Mistral integration updated to SDK v2
Notable features
  • ImageGeneration auto-fallback to subagent
  • agent field in RunContext
View release on GitHub
v1.75.0 Mixed
Notable features
  • gemini-embedding-2-preview types and limits
  • VertexAI Flex PayGo support
View release on GitHub
v1.74.0 Mixed
Breaking changes
  • Dataset without name parameter deprecated
Notable features
  • Online evaluation infrastructure
  • TextContent for metadata-only user prompts
  • Time-sortable run IDs with span attributes
View release on GitHub
v1.73.0 New feature
Notable features
  • CaseLifecycle hooks for Dataset.evaluate
  • ModelRetry exception in hooks
  • Model swapping via ModelRequestContext
View release on GitHub
v1.72.0 Mixed
Notable features
  • anthropic_eager_input_streaming to AnthropicModelSettings
  • Sync tool preparation functions
  • MCP URL parameter made optional
View release on GitHub
v1.71.0 New feature
Notable features
  • Capabilities composable behavior units
  • AgentSpec YAML/JSON loading with templated instructions
  • Hooks capability with decorators
View release on GitHub
v1.69.0 New feature
Notable features
  • Agent description parameter with span attribute
  • Direct multimodal tool results to APIs
  • Response-based fallback for FallbackModel
View release on GitHub
v1.67.0 New feature
Notable features
  • GPT-5.4 support
  • WebSearchTool for OpenRouterModel via plugins
  • Tavily search tool overhaul
View release on GitHub
v1.66.0 New feature
Notable features
  • Qwen 3.5 native structured output support
  • gemini-3.1-flash-image-preview model
View release on GitHub
v1.65.0 New feature
Notable features
  • UploadedFile object for provider files
  • gemini-3.1-flash-lite-preview model
View release on GitHub
v1.63.0 New feature
Notable features
  • Gemini 3.1 Pro Preview support
  • Logprob support for Gemini
  • args_validator parameter for tools
View release on GitHub
v1.62.0 New feature
Notable features
  • Tool approval integration for Vercel AI
  • LinePlot analysis type
  • ROCAUCEvaluator and KolmogorovSmirnovEvaluator
View release on GitHub
v1.60.0 Mixed
Notable features
  • video_url support to OpenRouterModel
  • instrumentation version 4 for OTel multimodal
View release on GitHub
v1.59.0 New feature
Notable features
  • Model.model_id property
  • opt-in flag for aggregated usage attribute names
  • enhance Contains evaluator for pydantic.BaseModel
View release on GitHub
v1.58.0 New feature
Notable features
  • report-level evaluators
  • multi-run aggregation support
  • extra_headers for Google provider
View release on GitHub
v1.56.0 Mixed
Security fixes
  • disallow downloading FileUrls pointing at local network by default
Notable features
  • Claude Opus 4.6 with anthropic_effort and adaptive thinking
  • anthropic_betas setting
View release on GitHub
v1.55.0 Bug fix

Fixed OpenAI compatible provider reasoning_tokens None handling and MultiModalContent union discriminator for deserialization

View release on GitHub
v1.52.0 Mixed
Notable features
  • openai_store setting for data retention
  • output validation retries in run context
  • Vercel AI SDK v5 compatibility
View release on GitHub
v1.50.0 Mixed
Notable features
  • usage_limits and model_settings exposure for to_cli()
  • OpenAI raw text annotations in TextPart
View release on GitHub
v1.49.0 New feature
Notable features
  • Parallel tool calls in DBOSAgent
  • BedrockEmbeddingModel for Nova, Cohere, Titan
View release on GitHub
v1.48.0 New feature
Notable features
  • allowed_domains for WebSearchTool
  • continuous_usage_stats model setting
View release on GitHub
v1.47.0 Bug fix

Ensured thought signatures and provider metadata survive round-trip through Vercel AI frontend.

View release on GitHub
v1.46.0 Mixed
Breaking changes
  • GrokProvider deprecated in favor of XaiModel
Notable features
  • XaiModel using xAI SDK
View release on GitHub
v1.45.0 New feature
Breaking changes
  • OpenRouter now sends DocumentUrls instead of downloading bytes
Notable features
  • VoyageAI embeddings support
View release on GitHub
v1.44.0 New feature
Notable features
  • Exa search tools integration
  • Bedrock Nova 2.0 Code Interpreter support
View release on GitHub
v1.42.0 New feature
Notable features
  • SambaNova provider support
  • OTel GenAI semantic convention attributes
View release on GitHub
v1.41.0 New feature
Notable features
  • YAML and TOML media type support
  • metadata for DeferredToolResults
View release on GitHub
v1.40.0 Mixed
Notable features
  • Temporal activity summaries
  • Tool call retry for nonexistent calls
View release on GitHub
v1.39.1 Bug fix

Fixed partial_output flag behavior in streaming output, prevented duplicated tool calls when using cache results, updated thoughtSignature for Vertex Gemini 3 compatibility, and corrected input_tokens calculation for Bedrock models.

View release on GitHub

Beta — feedback welcome: [email protected]