OpenClaw suddenly burned $0.13 in API costs – did I misconfigure something?
Setup:
Hardware: Raspberry Pi 4 (4GB RAM)
Software: OpenClaw + DeepSeek API
Models configured: Both DeepSeek-Chat (V3) and DeepSeek-R1
The issue:
I just finished setting up OpenClaw via onboard and applied a custom config JSON
Checked DeepSeek dashboard today and saw:
$0.13 USD consumed
1.82 million tokens
49 API requests
Average ~37k tokens per request
The problem: I never manually switched to R1
My openclaw.json clearly shows:
"primary": "deepseek/deepseek-chat"
I didn't run /model deepseek/deepseek-reasoner
I don't have any custom skills installed that might force R1
What I've checked:
Config file is correct (default model = V3)
No reasoner in my skills directory (grep -r found nothing)
Telegram logs show no accidental model switch commands
Questions:
Could OpenClaw be auto-routing complex tasks to R1 without telling me?
Is there any hidden skill or plugin that might override the default model?
How can I permanently disable R1 completely and only allow V3?
My config (relevant parts):
{
"models": {
"providers": {
"deepseek": {
"models": [
{"id": "deepseek-chat", "name": "DeepSeek Chat"},
{"id": "deepseek-reasoner", "name": "DeepSeek R1"}
]
}
}
},
"agents": {
"defaults": {
"model": {"primary": "deepseek/deepseek-chat"}
}
}
}
DeepSeek dashboard shows:
Consumption: $0.13
Tokens: 1.82M
Requests: 49
Average tokens/request: 37k
Any idea why R1 is being called when I never requested it? Is there a way to completely remove R1 from OpenClaw so it's not even an option?
Hardware: Raspberry Pi 4 (4GB RAM)
Software: OpenClaw + DeepSeek API
Models configured: Both DeepSeek-Chat (V3) and DeepSeek-R1
The issue:
I just finished setting up OpenClaw via onboard and applied a custom config JSON
Checked DeepSeek dashboard today and saw:
$0.13 USD consumed
1.82 million tokens
49 API requests
Average ~37k tokens per request
The problem: I never manually switched to R1
My openclaw.json clearly shows:
"primary": "deepseek/deepseek-chat"
I didn't run /model deepseek/deepseek-reasoner
I don't have any custom skills installed that might force R1
What I've checked:
Config file is correct (default model = V3) No reasoner in my skills directory (grep -r found nothing) Telegram logs show no accidental model switch commandsQuestions:
Could OpenClaw be auto-routing complex tasks to R1 without telling me?
Is there any hidden skill or plugin that might override the default model?
How can I permanently disable R1 completely and only allow V3?
My config (relevant parts):
{
"models": {
"providers": {
"deepseek": {
"models": [
{"id": "deepseek-chat", "name": "DeepSeek Chat"},
{"id": "deepseek-reasoner", "name": "DeepSeek R1"}
]
}
}
},
"agents": {
"defaults": {
"model": {"primary": "deepseek/deepseek-chat"}
}
}
}
DeepSeek dashboard shows:
Consumption: $0.13
Tokens: 1.82M
Requests: 49
Average tokens/request: 37k
Any idea why R1 is being called when I never requested it? Is there a way to completely remove R1 from OpenClaw so it's not even an option?
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
134,827Members
Resources
Similar Threads
Was this page helpful?
