How Third Moment Research Runs 10,000 Agents for 36+ Hours with OpenAI Codex and GPT-6
TL;DR
Third Moment Research runs 10,000 agents simultaneously for investigations lasting 36–48 hours, using OpenAI Codex, GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol. The system is a self-improving AI loop: models analyze agent behavior, identify recurring weaknesses, generate synthetic scenarios around those problems, and improve agents as new issues emerge. Daybreak automates security testing of the environments where those agents work.
What Third Moment Research built
Third Moment Research is building a self-improving AI system designed around the idea that the best agents should be very good at a company’s specific workflows, tools, edge cases, and definitions of success, and should improve the more they are used.
The company uses OpenAI models to understand agent behavior, identify recurring weaknesses, and automatically create thousands of synthetic scenarios around those problems. Third Moment Research then uses those scenarios to improve its agents, repeating the process as new issues emerge.
Codex, GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol support the system’s research and computational work. That combination enables Third Moment Research to run 10,000 agents at once on investigations lasting 36–48 hours. The company also uses Daybreak to test the security of the environments where agents work.
“In a single session, GPT-6.1 Sol is able to maintain a coherent thread of computational research across days of compute without losing the plot.”
—Adam Khakhar, Co-founder, Third Moment Research
Codex cut backtest runtime from about 3 hours to about 3–5 minutes
One of Third Moment Research’s early applications is quantitative finance, a domain with a clear measure of success. The team uses Codex to find and address problems in the research environment.
In one instance, Third Moment Research used Codex with GPT-5.6, the latest GPT model at the time, to analyze thousands of recorded simulation trajectories — the agents’ actions and the environment’s responses — alongside timestamps and the codebase.
Codex uncovered a bottleneck: backtesting a trading strategy across a year of order-level US equity data took several hours. While waiting, agents repeatedly polled for completion, wasting tokens and filling their context with status checks.
After profiling the backtest and analyzing its inputs and usage, Codex identified a caching optimization and a memory-packing technique. Codex experimented with solutions and deployed the fix to production, reducing the backtest from approximately 3 hours to approximately 3 to 5 minutes. That reduction lowered token use and context pollution, freeing more of the agents’ token budget for further experiments.
Agent harness: Codex inside a Dockerized environment
Third Moment Research gives each agent the tools, information, and environment it needs for a specific job. For quantitative research, the team runs the default Codex harness inside a Dockerized environment. Agents receive a terabyte of data, documentation, and a research repository with modules for feature engineering, GPU-supported model training, trading strategies, and simulation.
GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol provide the reasoning. The Codex harness gives agents tools to act, create plans, set objectives, spawn subagents, and compact context as work continues.
To address agents that stop before completing a task, Third Moment Research treats Codex as a subagent within an outer harness powered by GPT-6 Sol or GPT-6.1 Sol. The outer harness examines the trajectory when a task is left unfinished and determines whether to retry, terminate the job, or escalate to human review.
“The Codex harness is extremely versatile, and it enables use across many different domains.”
—Adam Khakhar, Co-founder, Third Moment Research
In that research environment, agents work to improve a baseline trading system built by Third Moment Research’s engineers. Agents test trading strategies on historical data, assess whether an improvement is meaningful or noise, and use each result to choose the next experiment. According to Adam, their work has improved the baseline trading system.
Model progression: from prescribed experiments to root-cause investigation
Before GPT-5.5, Adam says models could only execute experiments prescribed by a user. With GPT-5.5 and GPT-5.6, models working with the Codex harness could use results to design follow-up experiments. With GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol, agents can investigate root causes through hundreds of experiments running in parallel with subagents.
That progression has helped agents explore research directions the team had not anticipated. In one session, GPT-6.1 Sol identified the impact of non-displayed trades on an exchange where those trades accounted for an unexpectedly large share of volume.
“This was an unobvious result that our research team did not initially consider.”
—Adam Khakhar, Co-founder, Third Moment Research
Model evaluation and selection
Third Moment Research evaluates models across optimization, machine learning systems, and quantitative research. It tests frontier releases at their highest reasoning settings, assessing their ability to navigate repositories, write efficient code, and sustain an investigation. Current testing also includes GPT-6 Luna.
“GPT-6.1 Sol is currently our most-used model given its token efficiency. It completes tasks with the fewest tokens while delivering performance similar to GPT-6 Astra.”
—Jeffrey Yan, Co-founder, Third Moment Research
One benchmark asks agents to reduce a live scoreboard service’s update latency while preserving its outputs. The team measures speedup by dividing the average runtime of an expert-crafted solution by that of the model’s solution on the same hardware, subject to correctness. GPT-6 Astra delivered the biggest speedup, according to Adam.
For targeted work, including checks for specific security weaknesses in the simulation system, the team uses more efficient models like GPT-5.6 Terra. For deep research, Adam believes stronger reasoning can justify higher cost: “In the space of quantitative finance, the value of finding some interesting signal is extremely high.”
Daybreak automates security testing at scale
With 10,000 agents running autonomously, Third Moment Research needs a way to test the security of its research environments at scale. The team uses the Daybreak cybersecurity platform to automate that testing before simulations launch.
The tests check whether an agent could break out of its research environment and access the wider network.
“By using Daybreak, we’re able to have agents themselves do penetration testing on the simulation and test the infrastructure and security of it.”
—Adam Khakhar, Co-founder, Third Moment Research
“Without Daybreak, we would not be able to autonomously test this at scale.”
—Adam Khakhar, Co-founder, Third Moment Research
Roadmap: hundreds of thousands of collaborating subagents
Third Moment Research’s ambition extends to the specialized work companies do every day. For a retailer, that could mean making better supply chain decisions when inventory is short or a supplier is delayed. For an engineering team, it could mean finding performance bottlenecks in a large codebase. For a document-heavy business, it could mean identifying where commitments in one document do not match outcomes recorded in another.
Building on the reasoning capabilities of GPT-6 Astra, GPT-6 Sol, and GPT-6.1 Sol, Third Moment Research plans to expand from individual investigations to hundreds of thousands of collaborating subagents. Its roadmap calls for trillions of tokens supporting chains of dozens of dependent experiments, with the goal of using that research to autonomously improve agents as new challenges arise.
For Adam, this would give people more capacity to shape how agents learn and work:
“Humans can be in a much more leveraged position, where they can think about how to do research and optimize the research itself instead of actually doing the research.”
—Adam Khakhar, Co-founder, Third Moment Research