Hi team,
We are not able to see the thinking tokens/Reasoning tokens in the traces. Please do let us know, how can we get this information.
There’s no setting for it the field just doesn’t exist yet. CompletionUsage in llm.pyagents/livekit-agents/livekit/agents/llm/llm.py at main · livekit/agents · GitHub ) only carries prompt / completion / cached / total, and trace_ types.py (agents/livekit-agents/livekit/agents/telemetry/trace_types.py at main · livekit/agents · GitHub ) only defines gen_ai.usage. attributes for those. reasoning_tokens returns zero hits anywhere in the repo, so the span has nothing to emit.
You’re not losing the count though both OpenAI and Anthropic report reasoning/thinking tokens as part of output, so they’re already inside your completion_tokens. To break them out you’d have to read the raw provider usage yourself and set your own span attribute. Worth opening a feature request I couldn’t find one already filed.
This will not work with Gemini since it is emitting seperately
Not the answer you were looking for I’m afraid, but I see this GH issue includes adding thought summaries to OTEL but was closed automatically as not planned:
opened 04:47PM - 06 Dec 25 UTC
closed 12:13AM - 05 Jul 26 UTC
bug
### Bug Description
Gemini’s API provides thought summaries when include_though… ts=true is enabled. These appear in content.parts with the part.thought flag set and must be handled separately from normal answer text. [https://ai.google.dev/gemini-api/docs/thinking#summaries](<https://ai.google.dev/gemini-api/docs/thinking#summaries>)
Right now, the Google LLM plugin in `agents/livekit-plugins/livekit-plugins-google/livekit/plugins/google /llm.py` ignores part.thought entirely. The parsing code only looks at part.function_call and part.text:
```python
def _parse_part(self, id: str, part: types.Part) -> llm.ChatChunk | None:
if part.function_call:
chat_chunk = llm.ChatChunk(
id=id,
delta=llm.ChoiceDelta(
role="assistant",
tool_calls=[
llm.FunctionToolCall(
arguments=json.dumps(part.function_call.args),
name=part.function_call.name,
call_id=part.function_call.id or utils.shortuuid("function_call_"),
)
],
content=part.text,
),
)
return chat_chunk
return llm.ChatChunk(
id=id,
delta=llm.ChoiceDelta(content=part.text, role="assistant"),
)
```
There’s no check for part.thought, so thought summaries are treated as regular assistant output.
### Problems
1. Bug: Thoughts are spoken by TTS<br>When include_thoughts=True, thought summaries are merged into the same content used for user-facing responses. The TTS layer receives them and reads the agent’s internal reasoning out loud, which is not what Gemini’s “thinking” feature is meant for.
2. Missing observability of thoughts in OTEL<br>LiveKit already uses OpenTelemetry, but the Gemini thought summaries are not surfaced there at all. There is no way to inspect the model’s internal reasoning in traces/logs while keeping it hidden from the end user and TTS.
### Expected Behavior
**Separation of thoughts vs. answer**<br>Parts with part.thought == True should not be included in the assistant’s user-visible content.<br>Thought parts should never be forwarded to TTS or any channel that is meant for end-user output.
**Observability via existing OpenTelemetry**<br>Thought summaries should be attached to the existing OTEL spans/traces for Gemini calls
### Reproduction Steps
```bash
e.g.:
session = AgentSession(
stt="assemblyai/universal-streaming-multilingual",
llm=google.LLM(
model="gemini-2.5-flash-preview-09-2025",
temperature=0.8,
thinking_config=types.ThinkingConfig(
include_thoughts=True,
thinking_budget=1500,
),
),
tts=elevenlabs.TTS(
voice_id=elevenlabs_voice_id,
model=elevenlabs_model,
language=tts_language,
),
)
```
### Operating System
macOS Tahoe
### Models Used
assemblyai, google plugin, eleven labs
### Package Versions
```bash
# Core LiveKit dependencies
livekit>=1.0.13
livekit-agents[images,elevenlabs]>=1.3.6
livekit-api>=1.0.5
livekit-protocol>=1.0.6
# LiveKit plugins
livekit-plugins-google>=1.3.6
# Google Gemini API
google-generativeai==0.8.3
# Telemetry (Langfuse/OpenTelemetry/Judgment Labs)
opentelemetry-api>=1.39.0
opentelemetry-sdk>=1.39.0
opentelemetry-exporter-otlp-proto-http>=1.39.0
judgeval>=0.1.0 # Judgment Labs tracing (OpenTelemetry compatible)
```
### Session/Room/Call IDs
*No response*
### Proposed Solution
```python
```
### Additional Context
*No response*
### Screenshots and Recordings
*No response*