oracle-researchclaw-hybrid-model-opus-haiku-not-gpt
AutoResearchClaw’s internal 23-stage pipeline should use Opus 4.6 for the 8 critical reasoning stages (hypothesis generation, paper writing, peer review, literature analysis) and Haiku 4.5 for the 15 routine stages (topic init, resource planning, export, status checks). GPT-4o-mini was initially proposed but rejected in favor of Haiku 4.5 to stay within the Anthropic ecosystem and avoid external API dependencies. AutoResearchClaw natively supports primary + fallback model configuration.