Your hardware is custom.
Your software is custom.
Your agents should be too.
Our reported benchmark results show 31.3% higher bug-finding performance and 27% less debugging time with HyperBench compared to Codex or Claude.
At HyperBench, we’re building engineering agents with context tailored to the systems they work on. Our system graph connects source code, hardware specifications, configuration, and interface requirements, while custom retrieval brings the relevant evidence into each investigation.
We compared six models with and without HyperBench to examine how that context affects fault diagnosis and debugging.
We left OpenAI and Anduril to build this.
We gave AI agents a real drone and a test bench -- and they out performed vanilla Codex & Claude at finding and fixing bugs, with 20%+ gains on speed and correctness.
Check out how we did it:
We left OpenAI and Anduril to build this.
We gave AI agents a real drone and a test bench -- and they out performed vanilla Codex & Claude at finding and fixing bugs, with 20%+ gains on speed and correctness.
Check out how we did it:
AI agents struggle with engineering tasks that require reasoning across hardware, firmware, and software boundaries. Research such as HSCO-Bench (lnkd.in/g7r-wd-y) exposes failures at those boundaries, where the hardware, software interfaces, and configuration must stay consistent.
At HyperBench, we’re building agents that can follow relationships across the system to investigate failures and identify repairs. Our system graph connects source code, hardware specifications, configuration, and interface requirements, helping agents find the evidence needed to understand what went wrong.
We evaluated our system across 86 diagnostic and retrieval questions and 23 debugging cases. Some highlights:
- Up to 𝟑𝟏.𝟑% 𝐡𝐢𝐠𝐡𝐞𝐫 reported bug-finding performance over the base model*
- 𝟐𝟒.𝟑% 𝐥𝐨𝐰𝐞𝐫 median debugging time across repair/debugging tasks over the base model*
- 𝟐.𝟏𝟐× mean evidence recall: 36.3% → 77.0% across 50 hardware-retrieval development questions
Engineers reconstruct these relationships whenever they investigate a complex failure. We’re making that system knowledge accessible to agents throughout the investigation, from finding relevant evidence to identifying the implementation and requirements a repair needs to preserve.
Read the full blog here: hyperbench.ai/blog/system-gr…
*base model = GPT 5.5
91 Followers 135 Followingsold a hoa docs startup i built while buying a home at 22. now building @builderscv. @stanford cs alum.
dm me for startup roles
148 Followers 5K FollowingBuilding Synthetic Institutions. Researcher by nature and curiosity drives me. Small angel to 50+ companies. My favorite activity is calling out BS.
5K Followers 49 FollowingSee The Build. Dirac’s BuildOS is the first AI Process Design Platform, unifying Engineering and Manufacturing using AI + CAD.
1K Followers 194 FollowingWe are an early stage venture capital firm that partners with entrepreneurs building companies that support multi-dimensional mobility.
167 Followers 4 FollowingThe marketplace for buying & selling compute.
In San Francisco on Nov 17.
Request your invite by Fri Oct 23:
https://t.co/n95Um3NDSw
12K Followers 644 FollowingPresident @ Allen Control Systems. Founder and CEO of Bbot, acq by DoorDash. Host of The Drone Ultimatum - number 1 defense tech podcast.
567 Followers 2 FollowingA subsidiary of Hanwha Aerospace, Hanwha Defense USA delivers proven platforms with a demonstrated process maturity successfully projected around the world.
11K Followers 443 FollowingFinTech by day. At night I build DIY robots that think with their bodies. Embodied AI, microcontrollers & open-source hardware for makers. DM for collab.
96K Followers 6K FollowingTechno-industrialist, believer in hard power and Pax Americana. Automated Factories and the New American Workforce are how we win.
686K Followers 257 FollowingOfficial X of The Boeing Company. Dedicated to advancing aerospace technology by investing in our people and encouraging innovation at every turn.
911K Followers 1K FollowingWe hope you have a pleasant journey with us, as we share the latest news and views on our family of aircraft, sustainable aviation and much more. ✈️