Google DeepMind says four Gemini 4 Argon agents beat Minecraft without any human help in 27 hours and 14 minutes, costing $1,620 at Argon's planned 2027 price
The DecoderMatthias Bastian
Context & Ripple Effects
Google DeepMind had already moved Gemini 4 Argon from a limited cybersecurity-partner release to the paid Gemini API and consumer tiers, while Artificial Analysis placed its high setting alongside GPT-6 Astra on an intelligence index. The Minecraft run supplies a more operational test: agents had to divide work, retain shared notes and recover from errors over a long sequence rather than answer isolated prompts.
It also extends a line of DeepMind work in which Gemini's long context supported robots completing instructed tasks. In public reaction, observers characterized the run as an accessible example of agent coordination; DeepMind said the agents communicated through in-game chat and wrote 2,783 programs.
First-order effects
Google DeepMind gains a costed demonstration that four Argon agents can coordinate code-driven work over 27 hours without human intervention, giving its API and enterprise offerings a concrete capability reference beyond benchmark scores.
Argon's planned 2027 price makes task-level spending part of the result: prospective users can compare a $1,620 end-to-end run with the value of a similarly bounded automation task.
Second-order effects
Model providers competing for coding and knowledge-work deployments, including the GPT-6 Astra comparison set, face pressure to show not only single-model scores but multi-agent reliability, coordination and total-run cost.
Buyers evaluating agent systems will put greater weight on orchestration overhead—agent communication, debugging and persistence—because those components determine whether a capable model completes a lengthy workflow.
Third-order effects
If repeatable across useful environments, agent evaluation will shift from prompt-level benchmarks toward auditable cost per completed workflow, linking model quality more tightly to inference and coordination economics.
The relevant product boundary may move from a standalone model API to a managed multi-agent stack, where providers compete on runtime tooling and dependable task completion as well as model intelligence.
The trend: Frontier-model competition is broadening from benchmark performance to the cost and reliability of coordinated agents completing long-horizon tasks.
Four Gemini 4 Argon agents beat Minecraft from scratch in 27 hours. Operating in a sandbox using only code and text, they wrote 2,783 programs, divided up jobs over chat, shared notes, and debugged errors together - all with no human help ↓ https://x.com/...
The paper on how a team of Gemini 4 agents beat Minecraft is worth a look. It is accessible example (partially because it is AI written) of spontaneous coordination among agents to accomplish a complicated goal & how their approach differs from humans/ https://arxiv.org/... [imag…
Four Gemini 4 Argon agents teamed up in Minecraft to defeat the Ender Dragon in 27 hours, writing every move as code and coordinating in the game's chat. Great work led by @maxencefaldor with Satinder Singh. Read more in our technical report: https://arxiv.org/....
An independent research experiment, not an official Minecraft milestone, and not approved by or associated with Mojang or Microsoft.Visuals use the Pixel Perfection pack by XSSheep & Nova_Wostra (CC BY-SA 4.0).
A groundbreaking Frontier result! @maxencefaldor and @robertarail did a phenomenal job in showing the Argon capabilities so effectively. What a model and what a team in every sense of the word!
Gemini 4 Argon agents in Minecraft collaborating, demonstrating remarkable long-term planning and problem-solving all the way to the end! Fascinating breakthrough. Congratulations to @maxencefaldor, @robertarail and Satinder Singh!
Congratulations to @maxencefaldor, @robertarail, Satinder Singh and the rest of the team! Four Gemini 4 Argon agents teamed up in Minecraft, writing code and sharing learnings for 27 hours to beat the Ender Dragon. A fantastic example of Argon's ability to plan, adapt, and solve …