OpenRouter Launches Fusion Tool That Combines Multiple AI Models Into One API Call, Outperforming Solo Models in Benchmarks
Summary
OpenRouter launches Fusion, a groundbreaking tool that combines multiple AI models into a single API call, with benchmark tests showing fused model panels outperforming every solo model — and a budget panel even beating GPT-5.5 at half the cost.
Key Points
- OpenRouter launches Fusion, a new tool that synthesizes outputs from multiple AI models into a single API call, using a judge model to analyze consensus points, contradictions, and unique insights across panel responses.
- Benchmark testing on 100 DRACO deep research tasks reveals that model panels consistently outperform individual models, with Fable 5 and GPT-5.5 fused together scoring 69.0%, surpassing every solo model, while a budget panel of Gemini 3 Flash, Kimi K2.6, and DeepSeek V4 Pro beats GPT-5.5 at half the cost.
- Fusion is accessible via a simple model slug swap, a chatroom interface, or customizable API plugins, and even running the same model twice in a panel yields a significant performance boost, suggesting the synthesis step itself drives meaningful gains beyond just model diversity.