🐛 Bug Report: `gen_ai.request.model` set to `unknown` for LangGraph/LangChain

Which component is this bug for?

Langchain Instrumentation

📜 Description

Both gen_ai.request.model and gen_ai.response.model are set to unknown for Spans generated by Tool invocation tasks orchestrated by LangGraph.

Sample Code

llm = init_chat_model(
    "us.anthropic.claude-3-5-haiku-20241022-v1:0",
    model_provider="bedrock_converse",
)

## Define search tool
from langchain_community.tools import DuckDuckGoSearchRun
search = DuckDuckGoSearchRun()
tools = [search]
llm_with_tools = llm.bind_tools(tools)

Sample Spans

{
    "name": "ChatBedrockConverse.chat",
    "context": {
        "trace_id": "0x686db07fa79973ed8f91c2890df88f62",
        "span_id": "0x03b083f096667e17",
        "trace_state": "[]"
    },
    "kind": "SpanKind.CLIENT",
    "parent_id": "0x43105f4fbe80e5d7",
    "start_time": "2025-07-08T23:57:54.893220Z",
    "end_time": "2025-07-08T23:57:57.336534Z",
    "status": {
        "status_code": "UNSET"
    },
    "attributes": {
        "traceloop.workflow.name": "LangGraph",
        "traceloop.entity.path": "chatbot",
        "gen_ai.system": "Langchain",
        "llm.request.type": "chat",
        "gen_ai.request.model": "unknown",
        "gen_ai.response.model": "unknown",
        "llm.request.functions.0.name": "duckduckgo_search",
        "llm.request.functions.0.description": "A wrapper around DuckDuckGo Search. Useful for when you need to answer questions about current events. Input should be a search query.",
        "llm.request.functions.0.parameters": "{\"properties\": {\"query\": {\"description\": \"search query to look up\", \"type\": \"string\"}}, \"required\": [\"query\"], \"type\": \"object\"}",
        "gen_ai.prompt.0.role": "user",
        "gen_ai.prompt.0.content": "tell me a joke",
        "gen_ai.prompt.1.role": "assistant",
        "gen_ai.prompt.1.tool_calls.0.id": "tooluse_sfylHPX7SmGH58ACbxr19Q",
        "gen_ai.prompt.1.tool_calls.0.name": "duckduckgo_search",
        "gen_ai.prompt.1.tool_calls.0.arguments": "{\"query\": \"best clean jokes currently popular\"}",
        "gen_ai.prompt.2.role": "tool",
        "gen_ai.prompt.2.content": "Jun 17, 2025 \u00b7 Discover 135+ hilarious clean jokes perfect for adults! Find funny appropriate jokes, adult clean humor, and family-friendly laughs. Jun 21, 2025 \u00b7 Looking for funny, clever, clean jokes for adults that make you laugh without crossing any lines? Whether you\u2019re at a dinner party, happy hour, or just need something witty to \u2026 May 27, 2025 \u00b7 Discover our collection of hilarious, family-friendly jokes that prove clean humor packs the biggest punch! From clever wordplay and dad jokes to knock-knock classics and \u2026 Mar 12, 2025 \u00b7 Enjoy a collection of the best clean jokes perfect for all ages! From classic one-liners to kid-friendly chuckles, spread joy and laughter with these lighthearted, family-friendly jokes. By \u2026 Nov 15, 2024 \u00b7 These jokes from Ask Reddit are perfect for adults, kids, and everyone in between! 1. Just got attacked by 6 dwarves. Not Happy. 2. A priest, a minister, and a rabbit walk into a \u2026",
        "gen_ai.prompt.2.tool_call_id": "tooluse_sfylHPX7SmGH58ACbxr19Q",
        "gen_ai.completion.0.content": "Alright, here's a classic clean joke for you:\n\nWhy don't scientists trust atoms?\nBecause they make up everything! \n\n*Ba dum tss* \ud83d\ude04 It's a bit of science humor that plays on the double meaning of \"make up\" - both as creating something and as being untrustworthy. I hope it brought a smile to your face!\n\nWould you like to hear another joke?",
        "gen_ai.completion.0.role": "assistant",
        "gen_ai.usage.prompt_tokens": 704,
        "gen_ai.usage.completion_tokens": 96,
        "llm.usage.total_tokens": 800,
        "gen_ai.usage.cache_read_input_tokens": 0
    },
    "events": [],
    "links": [],
    "resource": {
        "attributes": {
            "telemetry.sdk.language": "python",
            "telemetry.sdk.name": "opentelemetry",
            "telemetry.sdk.version": "1.33.1",
            "service.name": "langgraph-min",
            "aws.log.group.names": "test_trace-min",
            "cloud.resource_id": "fastapi-app",
            "telemetry.auto.version": "0.1.3-aws",
            "aws.local.service": "langgraph-min",
            "aws.service.type": "gen_ai_agent"
        },
        "schema_url": ""
    }
}

👟 Reproduction steps

1. Add opentelemetry-instrumentation-langchain as a dependency to enable observability for LangGraph Agents.
2. Send a prompt to the Agent that triggers a tool invocation, interacting with a Claude model via AWS Bedrock.

👍 Expected behavior

OpenTelemetry span includes the gen_ai.request.model attribute populated with the correct model name.

👎 Actual Behavior with Screenshots

gen_ai.request.model attribute is populated with unknown

🤖 Python Version

3.10

📃 Provide any additional context for the Bug.

No response

👀 Have you spent some time to check if this bug has been raised before?

• I checked and didn't find similar issue

Are you willing to submit PR?

None

Start in the opentelemetry-instrumentation-langchain path that handles LangGraph tool-invocation spans, using the provided AWS Bedrock sample as the reproduction. Compare the emitted request and response model attributes with the model configured through init_chat_model. Done means the span reports the correct model name instead of unknown for this flow.