AI/ML●●●Banger
AI image models hallucinate history, we built a method to fix it it
Naive prompts hallucinate history; structured knowledge injection raises accuracy from 12.5% to 83.3%.
Big BrainWizardrySolve My Problem
MysticBirdie
122mo ago
Back to browse
GitHub Repository
Code3DBench: Single-Image to Executable Low-Poly 3D Code Generation
0 starsPython
Code3DBench – single-image to executable Three.js benchmark
by vladgl94·Jun 3, 2026·3 points·4 comments
AI Analysis
●●SolidNiche GemBig Brain
Benchmarks AI generating Three.js code from images, not just static mesh outputs.
Strengths
- •1012-object CC0 dataset provides substantial ground truth for consistent model evaluation.
- •Geometry metrics like Chamfer L2 and F-scores validate mesh quality accurately.
- •Fixed scaffold ensures generated code executes safely without custom environment setup.
Weaknesses
- •Narrow audience limits adoption compared to general LLM benchmarks like HELM.
- •No raw run traces included, making independent verification of results harder.
Category
Target Audience
AI researchers, 3D developers, ML engineers
Similar To
HELM · BigBench · EvalPlus
Similar Projects
AI/ML●●Solid
An Interactive Text to SQL Agent Benchmark
Interactive DuckDB-WASM benchmark beats static leaderboards for agentic SQL eval.
Big BrainNiche Gem
nl
102mo ago
Developer Tools●Mid
Automated AI Model Tester for Pollinations.ai
Daily CI/CD health checks for Pollinations.ai models, but anyone can do this with cron.
Ship It
osk111
212mo ago
AI/ML●●●Banger
Legal RAG Bench
Legal RAG benchmark revealing embedding quality > LLM choice by 19-point margin.
Big BrainNiche GemSolve My Problem
beowa
413mo ago
AI/ML●●Solid
Pokémon SVG Generation LLM Benchmark
Finally, a benchmark that uses Pokémon to test if models understand complex geometry.
Niche GemCrowd Pleaser
haxfenx
2020d ago
AI/ML●●●Banger
Lance – image/video generation and understanding in one model
Unified video and image model trained from scratch on just 128 GPUs.
WizardryBig Brain
cleardusk
641514d ago