Zip-NeRF
Anti-aliased grid-based neural radiance fields — mip-NeRF 360 quality at 24× the training speed.
About
Zip-NeRF is Google Research's answer to the aliasing problem in fast grid-based NeRFs. Methods like Instant NGP train in minutes but produce jaggies and missing detail because grids have no notion of scale. Zip-NeRF ports the anti-aliasing ideas from mip-NeRF 360 — reasoning about volumes along a cone instead of points along a ray — into the grid-based framework, plus some signal-processing tricks like multisampling and prefiltered feature lookups.
The paper's headline numbers are striking: 8–77% lower error than either mip-NeRF 360 or Instant NGP alone, while training about 24× faster than mip-NeRF 360. It appeared at ICCV 2023 and is widely cited. The awkward part is that Google never released official code or weights — what exists are community reimplementations (mostly JAX/PyTorch ports of the multinerf codebase), so getting it running takes some legwork.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
- Paper-reported (arXiv 2304.06706): 8–77% lower error than mip-NeRF 360 and Instant NGP baselines
- Same paper: ~24× faster training than mip-NeRF 360
- ICCV 2023 publication
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (TIER verified, metrics-v1). Evidence: ICCV 2023 paper from Google Research (arXiv 2304.06706) with clear quantitative claims vs mip-NeRF 360 and Instant NGP; project page live. Strengths: highly cited, peer-reviewed, strong benchmark claims. Docked for: (1) Google never released official code or weights — only community reimplementations exist; (2) speed/error numbers are author-reported. License: no weights released.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.0/10
- Maintenance · recency, releases, issue load
- 3.0/10
- Openness · license + docs
- 4.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · ICCV 2023 paper from Google Research (arXiv 2304.06706), submitted 2023-04-13
- · Paper reports 8-77% lower error than mip-NeRF 360 / Instant NGP and ~24x faster training than mip-NeRF 360
- · Google never released official code or weights — only community reimplementations (e.g. zipnerf-pytorch)