⛏️

Voyager (Minecraft)

Self-ReportedCurated

Open-ended Minecraft agent — 3.3× items, 15.3× tech tree vs prior SOTA.

NVIDIA Research (Voyager)· Operating since May 25, 2023· active
Follow
Curated from arXiv 2305.16291 — Voyager — not claimed by or endorsed by the organization. Metrics cited only as the source states. Absent metrics render as [unknown].

Recent activity

Version cuts and proof, newest first — the living track record.

  1. Artifact · Voyager paper published — arXiv 2305.162913y ago

Spec sheet

The benchmark fields — designed for comparison across teams.

Topology
Solo + Tools
Agent count
1
Platform
GPT-4 API (Mineflayer/Minecraft)
Runs on
GPT-4 API (Minecraft environment)
Industries
gamingresearch
Task kinds
open-ended-explorationskill-acquisitionembodied-reasoning
Trust tier
Self-Reported
Proof entries
1

Topology & roster

Solo + Tools

Solo-plus-tools. Single GPT-4 agent with three internal subsystems: curriculum generator, skill library, and code execution environment. The agent's behavior emerges from the interaction of these components, not multi-agent coordination.

System wiring

Typical Solo + Tools layout — schematic, not verified wiring
Typical role-level schematic — not verified wiringdirectswrites towrites toHuman operatorHumanoperatorHUMANGATEAgentAgentBUILDERTool ATool ARESOURCETool BTool BRESOURCE
Node details

Typical Solo + Tools layout — schematic, not verified wiring

HumanHuman operatorHuman gate
Tool
Human operator
Autonomy
Human-gated
Sends
  • directs → Agent
BuilderAgent
Tool
Agent
Autonomy
Runs autonomously
Sends
  • writes to → Tool A
  • writes to → Tool B
Receives
  • directs ← Human operator
ResourceTool A
Tool
Tool A
Autonomy
Runs autonomously
Receives
  • writes to ← Agent
ResourceTool B
Tool
Tool B
Autonomy
Runs autonomously
Receives
  • writes to ← Agent

How a typical Solo + Tools team handles a task

Typical Solo + Tools layout — schematic, not verified wiring

  1. Task arrives

    Human operator directs Agent.

  2. Agent builds the work

    Agent builds the work.

  3. The artifact lands

    The artifact lands in Tool A: Agent contributes via "writes to". The artifact lands in Tool B: Agent contributes via "writes to".

  4. Human holds the last word

    Human operator holds final approval.

Replicate a typical Solo + Tools setup

Typical Solo + Tools layout — schematic, not verified wiring

Ingredients

  • HumanHuman operator
  • BuilderAgent
  • ResourceTool A
  • ResourceTool B

Setup order

  1. 1.Provision the substrate: Tool A and Tool B.
  2. 2.Wire Agent: it receives "directs" from Human operator.
  3. 3.Declare the human gate: Human operator holds final approval.
⛏️
VoyagerExplorer

GPT-4 API (Minecraft environment)

GPT-4

Performance metrics

Windowed metrics with provenance. [unknown] means it was not tracked — an honest hole beats an invented figure.

Unique items vs SOTA
3.3
evidence-linked

3.3× more unique items obtained vs prior SOTA (DEPS). Source: arXiv 2305.16291 [evidence_linked]

as of May 25, 2023
Tech tree speed vs SOTA
15.3
evidence-linked

15.3× faster tech tree milestone completion vs prior SOTA (DEPS). Source: arXiv 2305.16291 [evidence_linked]

as of May 25, 2023
Exploration distance vs SOTA
2.3
evidence-linked

2.3× longer distances explored vs DEPS (prior SOTA). Source: arXiv 2305.16291 [evidence_linked]

as of May 25, 2023

Token economics

Cost transparency is part of the honesty architecture. [unknown] means it was not tracked — not that it is zero.

No cost metrics on record. Cost tracking is hard across runtimes; honest absence beats invented figures.

Blueprint

Operational DNA — why it works, how it was built, and how it is overseen. Not files for sale; knowledge of the design.

Why it works

Automatic curriculum keeps the agent in a productive challenge range — not too easy, not impossible. Skill library prevents re-learning already-discovered capabilities. Iterative prompting with execution feedback creates a tight edit-run-fix loop. The combination enables compound skill growth over long sessions.

How it was built

GPT-4 API with Mineflayer JavaScript API for Minecraft control. Curriculum generation uses GPT-4 with exploration state context. Skill library stores executable JavaScript programs indexed by natural language description. Iterative prompting executes code, captures errors and environment feedback, and re-prompts for correction.

Oversight model

No human-in-the-loop in evaluation. The agent operates autonomously for extended exploration sessions. The automatic curriculum is GPT-4-generated based on current state and past discoveries.

Proof (1)

The team's shared track record — tasks, incidents, lessons, milestones. Per-entry provenance tags are always visible.

  1. ArtifactMay 25, 2023evidence-linked

    Voyager paper published — arXiv 2305.16291

    3.3× more unique items, 2.3× longer distances, 15.3× faster tech tree milestones vs prior SOTA (DEPS). No fine-tuning required.

    https://arxiv.org/abs/2305.16291

Sign in to add a proof entry.

Sign in

Attestations (0)

Named third-party statements from people with first-hand experience. Attestations are what separates Peer-Attested from Evidence-Linked.

No attestations yet. Worked with this configuration or agent? Attest to it using the form below — attestations are named third-party statements and are what separates Peer-Attested from Evidence-Linked.

Sign in to attest to this team.

Sign in