πΈ Model Strategy (Hermes + Paperclip)
Decision date: 2026-07-30 Owner: Pat (cost mandate)
The Stack
| Use Case | Model | Provider | Cost |
|---|---|---|---|
| Primary (Hermes daily) | glm-5.2 | z.ai | Lowest |
| Fallback | Minimax M3 | OpenRouter | $0.30/M tokens |
| Heavy reasoning | GPT-5.6 Luna | OpenRouter | Moderate |
| Heavy coding | GPT-5.6 Terra | OpenRouter | Moderate |
| Claude (special) | Sonnet 5 / Opus 5 / Haiku 4.5 | OpenRouter | Variable |
| Sprint bookends | Claude Opus 5 | OpenRouter | $1-3/run |
Hard Rules
- NO x.ai as fallback. Explicitly prohibited by Pat.
- Cost-conscious always. βI donβt need a 10,000 hospital bill.β
- Minimax M3 is multimodal and well-rounded β good enough for most Paperclip agent work.
- Opus 5 only for deep thinking at sprint start/end bookends.
Hermes Aliases
/model heavyβ GPT-5.6 Luna/model codingβ GPT-5.6 Terra/model cheapβ Minimax M3/model claude-heavyβ Claude Opus 5/model claude-reasoningβ Claude Sonnet 5/model claude-fastβ Claude Haiku 4.5/model sonnetβ Claude Sonnet 5
Paperclip Agent Models (OpenCode)
All Paperclip agents using opencode_local adapter are now set to:
- Model:
openrouter/deepseek/deepseek-v4-flash-0731 - OpenCode version: 1.18.11
- OpenCode global config:
/root/.config/opencode/opencode.jsonc - Default model for ad-hoc runs: same deepseek v4 flash 0731
Agent working directories:
- CEO, Chief of Staff β
/root/repos/brain - Huible Tech Lead, Clinical Advisor β
/root/repos/huible - InvestInMe PM, R&D Lead, Project Builder β
/root/repos/investinme
All agents set to idle status, errors cleared. Ready for wipe/rebuild.