OpenRouter launches Fusion API, a compound model designed to match Fable's performance at half the cost

Notably, the budget panel was comparable with Claude Fable 5 in performance.

A panel of Gemini 3 Flash, Kimi K2.6, and DeepSeek V4 Pro, fused together, beat solo GPT-5.5 and solo Opus 4.8 outright.

And it landed within 1% of Fable 5 while costing roughly half the price.

How does it work?

When you send a prompt to Fusion, we fan it out to a panel of models in parallel, each with web search and bash tools enabled.

A judge model reads every response and extracts the structure: consensus points, contradictions, partial coverage, unique insights, blind spots.

Chatroom: https://openrouter.ai/fusion

@OpenRouter Looks like something Hermes Agent users might like!

Fusion is neurodiversity, but for models. Try it now!

💬 Chatroom: http://openrouter.ai/fusion (pick a preset or build a custom panel)

⚙️ API: docs at http://openrouter.ai/docs/guides/features/server-tools/fusion

ℹ️ More info on the blog post: https://openrouter.ai/blog/announcements/fusion-beats-frontier/

We benchmarked Fusion on 100 hard research tasks and found:

1. Panels of models consistently outperform individual models 2. Beyond-frontier performance can be achieved with frontier panels 3. Panels of budget models can surpass frontier models at a much lower cost

We ran it on the DRACO deep research benchmark by Perplexity: 100 deep research tasks across 10 domains, from law and medicine to finance and product comparison.

Each task is graded against ~39 weighted criteria, and wrong answers carry negative weight. (You can't bluff your way to a high score by being verbose.)

https://arxiv.org/abs/2602.11685

By testing different combinations of models, we found that roughly three quarters of the lift that Fusion provides comes from synthesis, and one quarter from diversity.

Then a synthesizer writes the final answer grounded in that analysis

Fusion runs server-side, so developers can call it exactly like a single model slug: "openrouter/fusion"

Or let the model decide when to reach for it by adding {"type": "openrouter:fusion"} to your tools array.

Want to customize the panel? Pass your own participant models and synthesizer:

@OpenRouter I believe the results!

Built an extension for VS Code that allows Claude and Codex to iterate back and forth and the results are mind blowing.

Note: we have only evaluated one deep research benchmark so far, which did not include long-horizon tasks.

Fable's long-horizon abilities were extremely impressive, and it calls for future work to benchmark both on long-horizon tasks.

Cool — we actually built something similar as an open-source skill using OpenCode CLI. Same idea: fan out to multiple models in parallel, judge the outputs, synthesize into one answer. All runs locally with a single bash script.

https://github.com/samirpatil2000/skills/blob/main/opencode-fusion/SKILL.md

Will share our benchmarks soon!

One detail we want to call out: when we first gave the panel web search, models started surfacing the DRACO rubric online.

We excluded those domains across every model with a one-line config change to the OpenRouter web search tool config, then re-ran everything. All published numbers come from the clean setup.

@YashasGunderia We ran these benchmarks earlier in the week, before Fable was taken down

@OpenRouter Alternatively, you can ask Codex, Gemini, Grok, and 400+ OpenRouter models (Qwen, Kimi, DeepSeek) for second opinions or arbiter-mediated consensus. One MCP server for Claude Code, Codex, Cursor, Kiro, OpenCode. Measures which models earn their seat.

https://github.com/antonbabenko/deliberation

@OpenRouter Here's a prompt you can give your agent so you can run this in Claude Code :) - It creates an alias 'or' that doesn't change your regular claude setup.

https://gist.github.com/JustSuperHuman/cda6981a62aa4fd2e2ba8120a717b348

@IkeVictor Cost comparisons were performed including cache hits

@PavelPenev8 @OpenRouter Oh hey, just saw this!

Do a search in VS Code, or even Google for WAT321, but also here's one of many links. MIT license, free.

https://open-vsx.org/extension/WillyDrucker/wat321

If you have any questions let me know. Needs both accounts with both Claude and Codex of course.

@OpenRouter Whomever is interested in other workflows like that, you might find this collection very interesting: https://github.com/av/harbor/blob/main/docs/5.2.3-Harbor-Boost-Modules.md

@OpenRouter so what you're saying is...