openai[patch],anthropic[patch]: Populate usage_metadata on invoke, add docs

docs/core_docs/.gitignore

@@ -55,8 +55,6 @@ docs/how_to/tools_prompting.md
 docs/how_to/tools_prompting.mdx
 docs/how_to/tools_builtin.md
 docs/how_to/tools_builtin.mdx
-docs/how_to/tool_calls_multi_modal.md
-docs/how_to/tool_calls_multi_modal.mdx
 docs/how_to/tool_calling.md
 docs/how_to/tool_calling.mdx
 docs/how_to/structured_output.md
@@ -107,6 +105,10 @@ docs/how_to/output_parser_fixing.md
 docs/how_to/output_parser_fixing.mdx
 docs/how_to/multiple_queries.md
 docs/how_to/multiple_queries.mdx
+docs/how_to/multimodal_prompts.md
+docs/how_to/multimodal_prompts.mdx
+docs/how_to/multimodal_inputs.md
+docs/how_to/multimodal_inputs.mdx
 docs/how_to/migrate_agent.md
 docs/how_to/migrate_agent.mdx
 docs/how_to/logprobs.md
@@ -135,6 +137,10 @@ docs/how_to/extraction_examples.md
 docs/how_to/extraction_examples.mdx
 docs/how_to/example_selectors.md
 docs/how_to/example_selectors.mdx
+docs/how_to/document_loader_markdown.md
+docs/how_to/document_loader_markdown.mdx
+docs/how_to/document_loader_html.md
+docs/how_to/document_loader_html.mdx
 docs/how_to/custom_tools.md
 docs/how_to/custom_tools.mdx
 docs/how_to/custom_callbacks.md
@@ -164,4 +170,6 @@ docs/how_to/binding.mdx
 docs/how_to/assign.md
 docs/how_to/assign.mdx
 docs/how_to/agent_executor.md
-docs/how_to/agent_executor.mdx
+docs/how_to/agent_executor.mdx
+docs/integrations/llms/mistral.md
+docs/integrations/llms/mistral.mdx

docs/core_docs/docs/how_to/chat_token_usage_tracking.mdx

@@ -14,14 +14,15 @@ This guide assumes familiarity with the following concepts:
 
 This notebook goes over how to track your token usage for specific calls.
 
-## Using `AIMessage.response_metadata`
+## Using `AIMessage.usage_metadata`
 
-A number of model providers return token usage information as part of the chat generation response. When available, this is included in the `AIMessage.response_metadata` field.
-Here's an example with OpenAI:
+A number of model providers return token usage information as part of the chat generation response. When available, this information will be included on the `AIMessage` objects produced by the corresponding model.
 
-import CodeBlock from "@theme/CodeBlock";
-import Example from "@examples/models/chat/token_usage_tracking.ts";
+LangChain `AIMessage` objects include a [`usage_metadata`](https://api.js.langchain.com/classes/langchain_core_messages.AIMessage.html#usage_metadata) attribute for supported providers. When populated, this attribute will be an object with standard keys (e.g., "input_tokens" and "output_tokens").
 
+#### OpenAI
+
+import CodeBlock from "@theme/CodeBlock";
 import IntegrationInstallTooltip from "@mdx_components/integration_install_tooltip.mdx";
 
 <IntegrationInstallTooltip></IntegrationInstallTooltip>
@@ -30,18 +31,52 @@ import IntegrationInstallTooltip from "@mdx_components/integration_install_toolt
 npm install @langchain/openai
 ```
 
-<CodeBlock language="typescript">{Example}</CodeBlock>
+import UsageMetadataExample from "@examples/models/chat/usage_metadata.ts";
 
-And here's an example with Anthropic:
+<CodeBlock language="typescript">{UsageMetadataExample}</CodeBlock>
 
-import AnthropicExample from "@examples/models/chat/token_usage_tracking_anthropic.ts";
+#### Anthropic
 
 ```bash npm2yarn
 npm install @langchain/anthropic
 ```
 
+import UsageMetadataExampleAnthropic from "@examples/models/chat/usage_metadata_anthropic.ts";
+
+<CodeBlock language="typescript">{UsageMetadataExampleAnthropic}</CodeBlock>
+
+## Using `AIMessage.response_metadata`
+
+A number of model providers return token usage information as part of the chat generation response. When available, this is included in the `AIMessage.response_metadata` field.
+
+#### OpenAI
+
+import Example from "@examples/models/chat/token_usage_tracking.ts";
+
+<CodeBlock language="typescript">{Example}</CodeBlock>
+
+#### Anthropic
+
+import AnthropicExample from "@examples/models/chat/token_usage_tracking_anthropic.ts";
+
 <CodeBlock language="typescript">{AnthropicExample}</CodeBlock>
 
+## Streaming
+
+Some providers support token count metadata in a streaming context.
+
+#### OpenAI
+
+For example, OpenAI will return a message chunk at the end of a stream with token usage information. This behavior is supported by `@langchain/openai` >= 0.1.0 and can be enabled by passing a `stream_options` parameter when making your call.
+
+:::info
+By default, the last message chunk in a stream will include a `finish_reason` in the message's `response_metadata` attribute. If we include token usage in streaming mode, an additional chunk containing usage metadata will be added to the end of the stream, such that `finish_reason` appears on the second to last message chunk.
+:::
+
+import OpenAIStreamTokens from "@examples/models/chat/integration_openai_stream_tokens.ts";
+
+<CodeBlock language="typescript">{OpenAIStreamTokens}</CodeBlock>
+
 ## Using callbacks
 
 You can also use the `handleLLMEnd` callback to get the full output from the LLM, including token usage for supported models.

docs/core_docs/docs/how_to/index.mdx

@@ -194,6 +194,7 @@ All of LangChain components can easily be extended to support your own versions.
 
 - [How to: pass multimodal data directly to models](/docs/how_to/multimodal_inputs/)
 - [How to: use multimodal prompts](/docs/how_to/multimodal_prompts/)
+- [How to: call tools with multimodal data](/docs/how_to/tool_calls_multimodal/)
 
 ## Use cases