autoresearchclaw-hybrid-llm-opus-haiku-split

AutoResearchClaw should use Opus 4.6 for its 8 critical reasoning stages (hypothesis generation, literature analysis, paper writing, peer review — stages 2, 7, 8, 9, 14, 15, 17, 18) and Haiku 4.5 for the 15 routine stages (topic init, resource planning, export). GPT-4o-mini was explicitly rejected in favor of Haiku 4.5 to stay within the Anthropic ecosystem. AutoResearchClaw natively supports primary + fallback model configs, so this is configured via YAML.