- OpenRouter launched Fusion, an API that combines multiple cheaper AI models to match the performance of expensive frontier models like Claude Fable 5.
- A budget panel of models achieved near-Fable 5 performance on the DRACO benchmark at roughly half the cost.
- The launch comes as a U.S. export control directive forced Anthropic to suspend access to Fable 5 and Mythos 5 for foreign nationals.
- Fusion uses a judge model to analyze panel responses and a synthesizer, like Claude Opus 4.8, to merge them into a single grounded answer.
- While not a complete replacement for all tasks, Fusion offers a viable alternative for deep research and complex planning at a lower price point.
OpenRouter launched a server-side API on June 12 called Fusion that challenges the dominance of expensive, single-model AI systems. This new product efficiently combines a panel of cheaper models to deliver performance that rivals premium offerings like Claude Fable 5 at a fraction of the cost.
The timing is fortuitous. A U.S. export control directive recently forced Anthropic to suspend its Fable 5 and Mythos 5 models for all foreign nationals. Consequently, OpenRouter quickly moved to promise “Fable-level intelligence at half the price” on X. The technique sends a prompt to multiple models in parallel, each equipped with web search and bash tools.
A judge model then extracts consensus points, contradictions, and blind spots from every response. Finally, a synthesizer writes the final answer grounded in that collective analysis. Tested on Perplexity’s DRACO benchmark, a cheap panel landed within 1% of Fable 5’s score at roughly half the cost.
Researchers like Andrew Trask called it “a way bigger deal than it seems.” However, skeptics cited issues with coding results and poor tool calling. Nonetheless, for deep research where contradictions matter, the fusion of multiple perspectives proves effective.
The charts make a clear point. On complex synthesis work, an expensive solo model is no longer the only path to strong results. A fused group of accessible models can deliver comparable performance with a much smaller bill.
✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.
