Skip to content

All models rejected with "context window below minimum 64,000 tokens" — Telegram completely down #24140

@fwends

Description

@fwends

Problem

This morning all models went down on Telegram, completely breaking the bot. Every user hit the same error:

Sorry, I encountered an error (ValueError).
Model MiniMax-M2.7 has a context window of 32,768 tokens, which is below the minimum 64,000 required by Hermes Agent.  Choose a model with at least 64K context, or set model.context_length in config.yaml to override.

The same error fired for kimi-k2.6 (also 32,768 context) on a cron job (facebook-groups-daily-scan):

ValueError: Model kimi-k2.6 has a context window of 32,768 tokens, which is below the minimum 64,000 required by Hermes Agent.

Impact

  • Telegram bot completely unresponsive for all users
  • Cron jobs also failing with the same error
  • Previously working models (MiniMax-M2.7, kimi-k2.6) suddenly rejected with no config change on our end

Environment

  • Hermes v0.13.0 (2026.5.7)
  • macOS
  • Telegram polling
  • Models affected: MiniMax-M2.7 (32,768 ctx), kimi-k2.6 (32,768 ctx)

Notes

  • Both models were working fine before this error appeared
  • No config changes made by user
  • Error message suggests setting model.context_length in config.yaml to override — but this was not needed before, suggesting Hermes introduced a new minimum context length requirement in a recent update

Metadata

Metadata

Assignees

No one assigned

    Labels

    P1High — major feature broken, no workaroundcomp/agentCore agent loop, run_agent.py, prompt builderprovider/kimiKimi / Moonshotprovider/minimaxMiniMax (Anthropic transport)type/bugSomething isn't working

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions