Hi,
Have setup grok as the provider with gpt-oss-120B and gpt-oss-20b as the selected models for us with a fresh installation of pi (the agentic coding harness). I’ve started a brand new session, with an empty folder, and just asked pi to “explain the project” hoping that pi with say that it is an empty project or something similar. However, I get the error that seems to indicate that I passed almost 32k tokens and I get a “Error: 413 Request too large for model openai/gpt-oss-20b in organization… service tier on_demand on tokens per minute (TPM): Limit 8000, Requested 33128, please reduce your message size and try again”.
I do not have a global AGENTS.md, do not have global or local SKILL.md files, and since I just launched pi without resuming any session, I’d have expected context to be very small, the system prompt that pi uses (quite small).
On groq dashboard, the logs shows the input & output token to be zero, for 6-7 attempts, but shows nothing to indicate what the failed queries were. Any idea, what might be wrong ?