Project HydraFusion: Frontier quality via multi-model orchestration

zatkin

This has got to be the worst project name I've seen all year.

jawns

I have a custom agent that I call my "G.S.D." (Get Stuff Done) agent. It is explicitly not a foreman, task routing, or an orchestrator agent. It has a bias toward direct action and is instructed to only delegate when necessary. I've found that this approach yields significantly faster results, without much of a quality trade-off, than an agent whose primary impulse is to delegate. In contrast, HydraFusion starts with a task routing step, then sequential planning, execution, and review stages. My guess is that this workflow is best for people who are prioritizing cost over speed for the same level of quality.

gopalv

> One model drafts a result, an independent read-only critic from a different model family reviews it Multiple model vendors is key here, the cascade pattern doesn't need it, but the critique pattern does. Last Nov, my team wrote a paper ("Team of Rivals") on the difference between using an OpenAI model to Critique an Anthropic model's output vs running a self-review agent loop on the same vendor. The ablations [1] proved that neither company alone was better than using both. The paper was a general response to "What does your company do that Anthropic can't?" but more so a demonstration of how to make something 90%+ good with models which eval at 60% or so (& Gas Town post unblocked our "this is a trade secret" argument about the paper). [1] - https://github.com/t3rmin4t0r/critique-evals

fnoef

They should probably focus on “project hydrastable infrastructure” instead

ValentineC

From the article: > In controlled offline evaluations, HydraFusion’s selective coding workflows matched or exceeded the evaluated Opus 5 baseline Opus 5 (in practice) is not a good baseline to compare against. I'm shocked that they chose to primarily compare against Opus 5 in all the article's charts. It's pretty disingenuous that they're claiming "frontier" quality, but didn't compare against Fable or Sol.

guybedo

i've been using adversarial critique and reviews for many planning, solution design and implementation steps inside workflows. It's so effective and helps catching so many design flaws, implementations misses etc ... that i'm wondering how people manage to build complex/large projects with agents without this kind of process. Well, i actually built this thing because i couldn't get good results so i had to find a way. I'm gonna open source the whole thing but it needs some cleanup, there's a basic landing page here https://kodfactory.com if anyone wants to be notified when it's released on github. Yeah i know, the world really needs another software factory :-)

Roark66

I'm questioning their results. It doesn't take much to beat the frontier in single benchmarks if one puts extra software between the model and the harness. This is also a reason why comparing "naked models" for which weights are available and frontier where providers can do whatever they want behind the scenes is unfair. Specifically, all it took to boost Qwen3.8-27B to get 10% more points on SWEbench Pro and Terminal Bench 2.0 with a proxy that has just these basics: - tweaks few decode settings like slightly higher temperature - detects when model gets stuck and tells it to "go on" - detects responses cut in the middle, empty responses that contain only reasoning, formats not passing verification etc and tells the model to "try better" And that is it. 10% more. I admit on a subset of tasks, but results are results, even on a subset.

ElFitz

Maybe I'm just 2026's Dropbox guy, but... yes, and?

swedishagentic

This looks very similar to what I've seen with a different harness - https://github.com/AMAP-ML/LongHorizon-Harness . It's critique pattern looks like another version of lh-harness without looping toward an overarching goal

mhmdfromkarak

that's crazy

hdz

If you’re a copilot cli user (we exist) you know that the auto mode leaves something to be desired. It only changes models at the start of a session or after compaction. Most dev tasks are now sub agent heavy and there is no model routing on the sub agents. I do notice that there is no mention of effort levels in the comparisons. 5.6 Luna Max is really good and really cheap. A 5.6 Sol high orchestrator with 5.6 Luna max is cheap, has frontier level performance, and is faster than Sol alone. This can be accomplished with simple agent instructions. Looking forward to running my own benchmarks on HydraFusion to see how it fares. Gone are the days of a single model doing all of the work it seems, unless the work requires no tool calls.