# THU-LYJ-Lab/T3Bench

T3Bench: Benchmarking Current Progress in Text-to-3D Generation

Repository: https://github.com/THU-LYJ-Lab/T3Bench
Canonical: https://ross.abutalabs.com/products/t3bench
Homepage: https://t3bench.com/
Language: Python
License Family: other
Topics: 3d, text-to-3d, diffusion, nerf
Last push: 2023-10-24T07:05:13+00:00

## Health v2 (maintenance only)
Score: 27/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 76
- inputs: {"age_days": 1064, "days_push": 1044, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1098, forks 11 (observed 2026-08-28T04:03:34.777931+00:00)

## What it is
T3Bench is the first comprehensive benchmark for text-to-3D generation, providing 300 text prompts across three complexity levels plus automatic quality and text-alignment metrics based on multi-view images and LLM evaluation. It includes evaluation pipelines built on ThreeStudio for testing methods like ProlificDreamer, Magic3D, Latent-NeRF, Fantasia3D, DreamFusion, and SJC.

## Use cases
- benchmark text-to-3d generation models
- evaluate quality of 3d assets generated from text prompts
- measure text-3d alignment automatically
- compare latentnerf magic3d fantasia3d dreamfusion sjc prolificdreamer
- download precomputed mesh results for text-to-3d methods
- research on diffusion-guided nerf generation

## When to choose
- you need standardized prompts and metrics to evaluate a text-to-3d method
- you want reproducible comparisons against published text-to-3d baselines
- you are researching multi-view consistency or text-3D alignment scoring

## When to avoid
- you need a text-to-3d generator rather than an evaluation benchmark
- you want a maintained production tool - the project has no license and limited updates
- your 3D generation method is not supported by the ThreeStudio-based pipeline

## Facets
- artifact type: dataset
- maturity: maintenance
- function: benchmarking, machine-learning
- domain: artificial-intelligence, computer-vision, graphics, machine-learning
- platform: python
- tags: text-to-3d, 3d-generation, diffusion-models, nerf, benchmark, evaluation-metrics, prompt-dataset, llm-evaluation, evaluation, gpu, linux

## Member repositories
- THU-LYJ-Lab/T3Bench (main) score 27

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:34.777931+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:46:13.869942+00:00, confidence not recorded.
  - readme: https://github.com/THU-LYJ-Lab/T3Bench (fetched 2026-08-28T04:03:34.777931+00:00, sha a081340eaa78)
  - homepage: https://t3bench.com/ (fetched 2026-08-29T12:49:44.212559+00:00, sha dea715507924)
- Data as of 2026-08-30T08:39:29.467469+00:00.