Skip to main content
GameDev.net gamedev.net

Benchmark News

The latest Benchmark coverage curated for game developers.

MeshSplatBench: A Unified Benchmark for Triangle-Based Neural Rendering

MeshSplatBench benchmarks triangle-based neural rendering all the way from native optimization to Unity deployment. The study shows that rasterizable output alone is not enough for production use: topology, engine compatibility, …

Research arXiv cs.GR · 3 weeks, 3 days ago
The GENEA Challenge 2026: A Large-Scale Disentangled Evaluation of Speech-Driven Gesture Generation on the Seamless Interaction Dataset

A new GENEA Challenge benchmark pushed speech-driven gesture generation through four large-scale human studies, with more than 23,000 votes from 869 testers. The results were sobering: motion realism improved on …

Research arXiv cs.GR · 1 month, 2 weeks ago
Clarify Before Executing: A Self-Evolving Agent for Resolving Intent Asymmetry in 3D Tool Orchestration

A new 3D agent called CLARE tackles a common production problem: users give vague prompts, but toolchains need exact parameters. It pauses to ask clarifying questions before running expensive 3D …

Research arXiv cs.GR · 2 months ago
Can AI Draw Science? A Benchmark for Evaluating Scientific Figure Generation by Text-to-Image and Multimodal Models

A new benchmark, SciDraw-Bench, tries to measure whether AI can generate scientific figures that are actually usable, not just visually plausible. It focuses on 32 structured tasks across eight figure …

Research arXiv cs.GR · 2 months, 3 weeks ago
3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code

A new benchmark, 3DCodeBench, tests whether VLM agents can turn text and image prompts into procedural 3D code that actually works in modeling software. For technical artists and tools programmers, …

Research arXiv cs.GR · 3 months, 3 weeks ago
CutVerse: A Compositional GUI Agents Benchmark for Media Post-Production Editing

A new benchmark shows how far GUI agents still have to go in real creative tools: on 186 long-horizon media-editing tasks across 7 pro apps, existing agents only hit 36.0% …

Research arXiv cs.GR · 4 months ago
PhysGaia: A Physics-Aware Benchmark with Multi-Body Interactions for Dynamic Novel View Synthesis

PhysGaia introduces a groundbreaking physics-aware benchmark for Dynamic Novel View Synthesis, focusing on multi-body interactions. This tool is crucial for developers aiming to enhance realism in simulations, as it offers …

Research arXiv cs.GR · 5 months, 3 weeks ago
HistCAD: A Constraint-Aware Parametric History-Based CAD Representation, Dataset, and Benchmark with Industrial Complexity

A new CAD benchmark puts editability, not just shape match, at the center of evaluation. HistCAD packages 170,236 executable parametric sequences with explicit constraints, STEP files, renders, and text labels. …

Research arXiv cs.GR · 7 months ago
WorldScore: A Unified Evaluation Benchmark for World Generation

The WorldScore benchmark introduces a unified evaluation standard for world generation in games. It features 3,000 test examples across various world types, focusing on controllability, quality, and dynamics. This benchmark …

Research arXiv cs.GR · 9 months, 3 weeks ago