Datum Foundry NineLABS

INDEPENDENT THINKING.
OBSERVABLE INTELLIGENCE.

MODEL GAUNTLET / COMPETITOR PROFILE

Opus 5.5.

Rank 1 of 4 · 464.2/500 across B01–B05 · Wildcards 200/200.

B01 / Mobile OS

All models on B01 ↗

TECHNICAL

97/100

Judged by qwen3.8-flash-next · ceiling 100

HUMAN · FUNCTIONALITY

100/100

Band: Good

FINAL

98.2/100

Not capped

Opus/B01 delivers a comprehensive mobile OS simulation ('Kestrel OS') with all 8 core domain features implemented across 16 modular files (3,696 LOC). Automated QA confirms 25/25 tests passing across Chromium, Firefox, and WebKit with zero console errors and zero external dependencies.

Time
1 h 12 min
Client / effort
Claude Code CLI
Finished
17 October 2026
Brief 1.0 · prompt SHA-256
a00e95f7a9243ba2b2a9046d9ff7225c565196138c18322f4a0c659241bf9755

B02 / Blender Mining Drone

All models on B02 ↗
Opus 5.5: main image for B02 Blender Mining Drone
Opus 5.5 · B02 Blender Mining Drone. Produced by the model during the challenge.

TECHNICAL

98.6/100

Judged by qwen3.8-flash-next · ceiling 100

HUMAN · VISUAL QUALITY

90/100

Band: Good

FINAL

95.2/100

Not capped

Strong, deterministic-verified Blender mining drone submission. The .blend and .glb are present, valid, editable, and structurally coherent: 50 mesh objects, 67,252 triangles, 0 degenerate triangles, 0 non-manifold objects, complete UVs/normals, unit transforms, single hierarchy root, semantic naming, 16 PBR materials, 6 embedded textures, and 38/38 articulated pivots at joint bounds. Four distinct rendered views and the reference image are preserved, with palette alignment verified.

Client / effort
Claude Code CLI (Windows), session 8ade1e60-80ea-4fba-91f6-3c4678e65858
Finished
18 October 2026
Brief 1.0 · prompt SHA-256
4d9e8edee96c488962673f7c21868a8cba3ea55012bbef6c9e733f505dd413b3

B03 / Blender to Godot

All models on B03 ↗

TECHNICAL

90/100

Judged by qwen3.8-flash-next · ceiling 100

HUMAN · PLAYABILITY

96/100

Band: Good

FINAL

92.4/100

Not capped

Strong B03 submission with authentic same-model B02 drone GLB, clean Godot 4.7 project, launchers, documentation, and a 50/50 automated gameplay test covering movement, collision, mining, delivery, win, and restart. Core critical functionality appears verified by deterministic telemetry and QA logs.

Time
25 min
Client / effort
Claude Code
Finished
18 October 2026
Brief 1.0 · prompt SHA-256
f00af20d56033bae25fc2d36a139ec5f56e08b50d99be07946ad19c06a5f5527

B04 / Prime London Estate Agent

All models on B04 ↗

TECHNICAL

86/100

Judged by qwen3.8-flash-next · ceiling 100

HUMAN · DESIGN & NAVIGATION

96.7/100

Band: Good

FINAL

90.3/100

Not capped

Exceptionally well-engineered static site with 212 output files, 7,400 LOC, comprehensive test infrastructure (92 unit + 48 e2e across 3 browsers), and zero external dependencies. The agency 'Vane & Daughters' has a defined brand system with three bundled open-licence typefaces. All required pages exist including 6 area guides. The main limitation is that visual design quality, photography art direction consistency, and floor plan accuracy require live browser inspection to fully confirm.

Time
1 h 50 min
Tokens
106,057,445
Client / effort
Claude Code / max
Finished
23 October 2026
Brief 2.0 · prompt SHA-256
b2229dc93c7a8d4ec7dda8712ff98e2f0c8cbc618181cd8cc2239ac458d1bc50

B05 / Oliver GPU Game

All models on B05 ↗

TECHNICAL

83/100

Judged by qwen3.8-flash-next · ceiling 100

HUMAN · PLAYABILITY

96/100

Band: Good

FINAL

88.2/100

Not capped

Strong, technically substantial B05 submission. It provides a self-contained local browser game, official Oliver asset provenance verified by SHA-256, a 4-rack GPU thermal/job simulation, unit/e2e evidence, an assets manifest, and a recorded build/QA trail. The main gaps are visual/audio/runtime verification: no screenshots, no captured audio, no frame-timing telemetry, and some lifecycle/HUD/animation claims require human playtest or viewport inspection.

Time
1 h 56 min
Client / effort
Claude Code / max (set by the user with /effort before the prompt)
Finished
18 October 2026
Brief 1.0 · prompt SHA-256
0ad15378b12960bfc136ed8c3115ccb3b570e41cf6f65dc361dee3ae0efa91a6

B06 / Wildcard One

All models on B06 ↗
Opus 5.5: main image for B06 Wildcard One
Opus 5.5 · B06 Wildcard One. Produced by the model during the challenge.

CREATIVE SCORE

100/100

Scored by the reviewer

Time
25 min
Tokens
11,615,373
Client / effort
Claude Code / medium
Finished
19 October 2026
Brief 1.0 · prompt SHA-256
827031e6b27e4e8d209db933a2e3a8166476216433fe01c2a1fd4083be4ebfbd

B07 / Wildcard Two

All models on B07 ↗

CREATIVE SCORE

100/100

Scored by the reviewer

Time
5 h 41 min
Tokens
79,800,237
Client / effort
Claude Code / max
Finished
20 October 2026
Brief 1.0 · prompt SHA-256
e2555fa49a4563e27a174b1bef88a31085c0295186d7a4093e1f9832bf5c97c5
← All competitors