anthropic/claude-fable-5.1
claude fable 5.1
Artifact score77/100
Compliance85%
Frozen testSunwell Sanctuary ↗Voxel Forge · standard · Voxel
Evidence
explorer.
Find the exact model work Proofrun has published. Every tile is one artifact or incident under one frozen condition—not a model-wide score.
Evidence records27
Models15
Frozen tests8
Published reports16
tencent/hy4-preview
hy4 preview
Artifact score65/100
Compliance90%
Frozen testTriage Desk ↗Product workflow · standard · Web
tencent/hy4-preview
hy4 preview
Artifact score—
Compliance—
Frozen testWeather Story ↗Data visualization · quick · Web
tencent/hy4-preview
hy4 preview
Artifact score61/100
Compliance80%
Frozen testVerdant Terminal ↗Voxel Forge · standard · Voxel
qwen/qwen3.8-flash
qwen3.8 flash
Artifact score76/100
Compliance100%
Frozen testTriage Desk ↗Product workflow · standard · Web
qwen/qwen3.8-flash
qwen3.8 flash
Artifact score19/100
Compliance0%
Frozen testTriage Desk ↗Product workflow · standard · Web
openai/gpt-5.6-terra
gpt 5.6 terra
Artifact score79/100
Compliance100%
Frozen testTriage Desk ↗Product workflow · standard · Web
stealth/ox-alpha
ox alpha
Artifact score56/100
Compliance100%
Frozen testOrbit Radio ↗Interface design · quick · Web
z-ai/glm-5.3
glm 5.3
Artifact score46/100
Compliance100%
Frozen testOrbit Radio ↗Interface design · quick · Web
google/gemini-3.7-flash
gemini 3.7 flash
Artifact score34/100
Compliance100%
Frozen testOrbit Radio ↗Interface design · quick · Web
z-ai/glm-5.3
glm 5.3
Artifact score—
Compliance—
Frozen testOrbit Radio ↗Interface design · quick · Web
google/gemini-3.7-flash
gemini 3.7 flash
Artifact score11/100
Compliance0%
Frozen testOrbit Radio ↗Interface design · quick · Web
openai/gpt-5.6-terra
gpt 5.6 terra
Artifact score78/100
Compliance100%
Frozen testOrbit Radio ↗Interface design · quick · Web
x-ai/grok-4.6
grok 4.6
Artifact score86/100
Compliance100%
Frozen testPrism Garden ↗Creative coding · quick · Web
x-ai/grok-4.6
grok 4.6
Artifact score8/100
Compliance0%
Frozen testPrism Garden ↗Creative coding · quick · Web
x-ai/grok-4.6
grok 4.6
Artifact score92/100
Compliance100%
Frozen testPrism Garden ↗Creative coding · quick · Web
x-ai/grok-4.6
grok 4.6
Artifact score77/100
Compliance100%
Frozen testPrism Garden ↗Creative coding · quick · Web
x-ai/grok-4.6
grok 4.6
Artifact score89/100
Compliance100%
Frozen testPrism Garden ↗Creative coding · quick · Web
x-ai/grok-4.6
grok 4.6
Artifact score90/100
Compliance100%
Frozen testSunwell Sanctuary ↗Voxel Forge · standard · Voxel
deepseek/deepseek-v4-flash-0731
deepseek v4 flash 0731
Artifact score92/100
Compliance100%
Frozen testWeather Story ↗Data visualization · quick · Web
tencent/hy3
hy3
Artifact score90/100
Compliance100%
Frozen testWeather Story ↗Data visualization · quick · Web
anthropic/claude-fable-5
claude fable 5
Artifact score74/100
Compliance79%
Frozen testSky Archive ↗Voxel Forge · standard · Voxel
upstage/solar-pro4
solar pro4
Artifact score85/100
Compliance100%
Frozen testPrism Garden ↗Creative coding · quick · Web
openai/gpt-5.6-sol
gpt 5.6 sol
Artifact score70/100
Compliance92%
Frozen testPrism Garden ↗Creative coding · quick · Web
anthropic/claude-opus-5
claude opus 5
Artifact score77/100
Compliance92%
Frozen testPrism Garden ↗Creative coding · quick · Web
anthropic/claude-sonnet-5
claude sonnet 5
Artifact score73/100
Compliance92%
Frozen testGravity Desk ↗Physics toy · quick · Web
openai/gpt-5.6-terra
gpt 5.6 terra
Artifact score87/100
Compliance92%
Frozen testGravity Desk ↗Physics toy · quick · Web
How to read this index
Browse observations, not standings.
A model may appear more than once because it produced more than one preserved artifact. Scores belong to those artifacts under their original rubrics; they are not averaged across unlike tests. Provider incidents and no-submission records stay visible without receiving invented zeroes.
Open the linked report for the prompt, route, budgets, attempt history, evidence, human rationale, and limitations behind any tile.