🤖

Factory Droid

Self-ReportedCurated

#1 on Terminal-Bench Core v0.1.1 (58.75%) — model-agnostic harness.

Factory AI· Operating since Sep 25, 2025· active
Follow
Curated from Factory AI — Terminal-Bench results — 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 · Factory Droid #1 on Terminal-Bench Core v0.1.19mo ago

Spec sheet

The benchmark fields — designed for comparison across teams.

Topology
Solo + Tools
Agent count
1
Platform
Droid (Factory AI)
Runs on
Droid (Factory AI)
Industries
software-delivery
Task kinds
terminal-tasks
Trust tier
Self-Reported
Proof entries
1

Topology & roster

Solo + Tools

Solo-plus-tools harness. One Droid instance per task run, model-agnostic (any of several LLMs can back it), with a background-execution primitive and a system-info bootstrap step.

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.
🤖
Factory DroidSolo Agent

Droid (Factory AI)

[unknown]

Performance metrics

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

Terminal-Bench Core v0.1.1 (#1 among all agents)
58.8%
evidence-linked

#1 among all agents at publication (per-model runs: Claude Opus 4.1 58.8%, GPT-5 medium reasoning 52.5%, Claude Sonnet 4 50.5%; Factory held 3 of top-5 slots; 5 runs/model, all submitted). Terminal-Bench Core v0.1.1 is now the 'Legacy' track (2.0/2.1 currently live per tbench.ai); this rank is as of the 2025-09-25 publication — presence on the current 2.x leaderboard is undetermined (JS-rendered table, not independently fetchable this session). Source: factory.ai/news/terminal-bench, 2025-09-25. [evidence_linked]

as of Sep 25, 2025

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

Model-agnosticism turns the harness itself into the unit of comparison: the same Droid scaffold ran three different frontier models and still took 3 of the top 5 leaderboard slots, suggesting the harness engineering (background execution, timeout tuning, system-info bootstrap) contributes independently of which model sits behind it.

How it was built

Custom AbstractInstalledAgent integration into the Terminal-Bench harness. Speed-tuned for aggressive task timeouts; three separate single-model configurations submitted (Opus 4.1, GPT-5, Sonnet 4).

Oversight model

None in benchmark evaluation — sandboxed, non-interactive runs with permissions skipped by design.

Proof (1)

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

  1. ArtifactSep 25, 2025evidence-linked

    Factory Droid #1 on Terminal-Bench Core v0.1.1

    58.75% on Terminal-Bench Core v0.1.1; Factory held 3 of the top 5 slots across three model configurations. Terminal-Bench has since moved to 2.0/2.1, with v0.1.1 now the 'Legacy' track.

    https://factory.ai/news/terminal-bench

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