# New Zealand teams publishing AI research.

URL: https://levystreet.com/compare/ai-research-teams-nz/
Updated: 2026-10-19

AI research in NZ: which teams publish open-source AI tools, benchmarks and papers, under what licence, and how to judge the claims. Checked 20 October 2026.

## Short answer

New Zealand teams publishing AI research or open-source AI work include Levy Street (Wellington), the University of Waikato (Hamilton, home of Weka and MOA), the University of Auckland, Victoria University of Wellington, Te Hiku Media (Kaitāia) and Earth Sciences New Zealand. Choose by what you need: reusable code with a clear licence, peer-reviewed methods, or a benchmark you can rerun on your own tasks.

- Waikato's Weka and MOA are open-source machine learning tools released under the GNU GPL.
- University of Auckland researchers built LIBRA, over 360,000 test cases of language-model bias in a New Zealand context.
- Te Hiku Media's Kaitiakitanga License needs permission for any use and bars unapproved commercial use.
- Choosing rule: check the licence, the test data and who can reproduce the result before relying on it.
- Levy Street publishes open-source code and applied research; the research has not been peer reviewed.

Most public AI research in New Zealand comes from universities, public science organisations and Māori-led language technologists, with a few companies publishing code. For a business, research is useful as code you can reuse, methods you can trust or benchmarks you can rerun on your own work.


Levy Street wrote this guide and is one of the teams in it, listed first because this is our site. Every other team's details come from its own website, checked on 20 October 2026. No organisation paid to be included. If something here is wrong or out of date, [tell us](/contact/) and we will fix it.


## New Zealand AI research teams compared

Levy Street first, then alphabetical. "Releases" means code, data or benchmarks the team itself publishes.

| Team | Based in | Research focus | Releases and outputs | Licence | Suits |
|---|---|---|---|---|---|
| Levy Street | Wellington | Applied AI engineering: agents, evaluation, models written as code | World of ClaudeCraft, SPAWN D, Tiny Shakespeare CLM; co-authored GTM-Bench | MIT; MIT or Apache-2.0; none | Teams taking research into production |
| AI Researchers Association (AIRA) | National | Represents New Zealand AI researchers | October 2026 strategy whitepaper; directory of research groups | Not applicable | Finding the right researcher |
| Earth Sciences New Zealand | Nationwide | AI for climate and environmental science | Generative AI climate downscaling study (Climate Dynamics) | No code or data release mentioned | Climate and hazard risk |
| Te Hiku Media | Kaitāia | Te reo Māori speech and language technology | NeurIPS 2024 benchmark paper; Papa Reo API | Kaitiakitanga License (permission required) | Te reo Māori technology, by agreement |
| Te Pūnaha Matatini | National network, hosted by the University of Auckland | Complex systems; one completed AI project | Project posts and outputs | Not stated | Complex-systems questions |
| University of Auckland | Auckland | Neuro-symbolic AI, industrial AI, machine learning | Strong AI Lab code; HA-ViD dataset; LIBRA bias benchmark | MIT and Apache-2.0 code; HA-ViD non-commercial | Collaborative research |
| University of Canterbury | Christchurch | Image processing, machine learning, computer vision | Project pages | No releases found | Image learning problems |
| University of Otago (CAIPP) | Dunedin | AI policy, regulation, ethics and governance | Policy research | Not applicable | Policy and governance |
| University of Waikato | Hamilton | Machine learning algorithms, data streams, environmental AI | Weka, MOA, MEKA, CapyMOA | GPL; CapyMOA BSD-3-Clause | Proven open-source ML tools, consultancy |
| Victoria University of Wellington | Wellington | Evolutionary computation, deep learning, interpretable AI | Peer-reviewed papers | No releases found | Research partnerships, commercial projects |

## Team profiles

### Levy Street

Levy Street is a Wellington AI engineering company whose [research](/research/) feeds the systems it builds. Its [open-source work](/research/open-source/) includes [World of ClaudeCraft](/portfolio/world-of-claudecraft/), an MIT-licensed browser MMO with a Gymnasium environment for training reinforcement-learning agents (2,303 GitHub stars and 60 contributors on 19 October 2026), and [SPAWN D](/portfolio/spawnd/), a pre-alpha control plane for coding agents under MIT or Apache-2.0. [Tiny Shakespeare CLM](/research/code-language-models/), a language model written as Python and improved by a Claude agent, has no licence and reports training-set scores only. Three team members co-authored [GTM-Bench](/portfolio/gtm-bench/), a Blackpearl Group benchmark, and [LucentBench](/research/lucentbench/) is described in the whitepaper of Lucent, a Levy Street venture, with no public dataset. None of these papers or benchmarks has been peer reviewed.

### AI Researchers Association (AIRA)

[AIRA](https://www.airesearchers.nz/about) is a not-for-profit incorporated society, set up in 2021 to support AI research in New Zealand. It published the third edition of its [strategy whitepaper](https://www.airesearchers.nz/site_files/28243/upload_files/AIRA_Whitepaper_2026.pdf) in October 2026 and keeps a directory of research groups, which helped us find several entries here.

### Earth Sciences New Zealand

Earth Sciences New Zealand, formed when [GNS Science and NIWA merged](https://www.earthsciences.nz/about-us) on 1 July 2025, announced on 20 October 2026 a [Climate Dynamics study](https://www.earthsciences.nz/news/new-zealand-led-research-at-climate-sciences-ai-frontier-unlocks-extreme-weather-insights) that used an in-house generative AI model to produce more than 15,000 simulated years of daily rainfall and temperature at 12 km resolution. The article mentions no code or data release.

### Te Hiku Media

[Te Hiku Media](https://tehiku.nz/te-hiku-tech/papa-reo/), a Kaitāia charitable media organisation belonging to five Far North iwi, leads Papa Reo, its language technology programme, with Dragonfly Data Science. Its team published a [te reo Māori benchmark paper](https://neurips.cc/virtual/2024/poster/97451) in the NeurIPS 2024 Datasets and Benchmarks Track; the paper does not release new assets. Its code is under the [Kaitiakitanga License](https://github.com/TeHikuMedia/Kaitiakitanga-License), which requires permission for any use and bars commercial use without an explicit grant.

### Te Pūnaha Matatini

[Te Pūnaha Matatini](https://www.tepunahamatatini.ac.nz/our-research/artificial-intelligence) is a Centre of Research Excellence for complex systems, hosted by the University of Auckland. Its AI project studied how AI interacts with the systems it is used to study, with the aim of more interpretable algorithms; the research listing marks it [completed](https://www.tepunahamatatini.ac.nz/our-research).

### University of Auckland

Auckland's [Strong AI Lab](https://www.ai.ac.nz/) works on neuro-symbolic AI and publishes [code](https://github.com/Strong-AI-Lab) under MIT and Apache-2.0, such as an emotion recognition toolkit. Its [Industrial AI Group](https://uoa-iai.github.io/) released the [HA-ViD video dataset](https://iai-hrc.github.io/ha-vid) under a non-commercial Creative Commons licence. School of Computer Science researchers, with a Massey University co-author, published [LIBRA](https://arxiv.org/abs/2502.01679), over 360,000 test cases for measuring language-model bias in a New Zealand context. The university's proposed Aotearoa Agentic AI Platform received [$250,000 of seed funding](https://www.auckland.ac.nz/en/news/2025/12/18/new-zealand-ai-research-proposal-reaches-first-stage.html) to develop a full proposal.

### University of Canterbury

Canterbury's [Spatial and Image Learning group](https://www.canterbury.ac.nz/research/about-uc-research/research-groups-and-centres/spatial-and-image-learning--sail--group) researches image processing and machine learning, develops prototypes for deployment and invites organisations with an image learning problem to get in touch. Its [AI research group page](https://www.canterbury.ac.nz/research/about-uc-research/research-groups-and-centres/artificial-intelligence-research-group) describes the group as newly formed and refers to a 2022 workshop.

### University of Otago (CAIPP)

Otago's [Centre for Artificial Intelligence and Public Policy](https://www.otago.ac.nz/caipp) researches AI policy, regulation, ethics and governance, with attention to policy options for New Zealand. It says it advises government ministries. This is policy research, not model building.

### University of Waikato

Waikato's [AI Institute](https://www.waikato.ac.nz/research/institutes-centres-entities/institutes/artificial-intelligence-institute/), Te Ipu o te Mahara, and its machine learning group maintain [Weka](https://ml.cms.waikato.ac.nz/weka/) (GPL), [MOA](https://github.com/Waikato/moa) for data streams (GPL-3.0) and [CapyMOA](https://capymoa.org/) (BSD-3-Clause). Waikato [says](https://www.waikato.ac.nz/research/institutes-centres-entities/institutes/artificial-intelligence-institute/research/software/) its tools have more than 10 million downloads, and the institute [offers](https://www.waikato.ac.nz/research/institutes-centres-entities/institutes/artificial-intelligence-institute/engagement/) research, advisory and consultancy services.

### Victoria University of Wellington

Te Herenga Waka's [Centre for Data Science and Artificial Intelligence](https://www.wgtn.ac.nz/cdsai) runs research projects and commercial work for clients, plus postgraduate supervision. Its [Evolutionary Computation and Machine Learning group](https://ecs.wgtn.ac.nz/Groups/ECRG/) publishes peer-reviewed papers on genetic programming, feature selection and text mining.

We also checked [AI New Zealand's "NZ AI Benchmark"](https://www.newzealand.ai/benchmark/), which is a self-assessment of an organisation's AI adoption, not a test of AI systems, so it is not listed.

## How to judge AI research before you rely on it

Treat a result as evidence only when you could check it yourself.


- The licence is stated, and it allows your use (commercial, modified, redistributed).
- Scores come from held-out data, not the data the system was tuned on.
- The test set and scoring code are public, or you can rerun the test on your own tasks.
- The paper says who ran the benchmark and whether they also built a competing system.
- It has been peer reviewed, or you know it has not.
- Any Māori data was used with consent and under the terms its kaitiaki set.


Our [LLM evaluation service](/services/llm-evaluation/) builds these tests on a client's own documents rather than relying on public scores.

## Where AI research claims go wrong

- **"Open" without a licence.** Code on GitHub with no licence is source-available: you can read it, not reuse it. Our Tiny Shakespeare CLM is in that category.
- **Training-set scores.** A score measured on the data a system was tuned on overstates what it will do on new work.
- **Benchmarks only the operator can run.** The [GTM-Bench paper](https://gtm-bench.ai/assets/gtm-bench-paper.pdf), which Levy Street team members co-authored, states that its production data environment "limits immediate reproducibility outside the benchmark operator", and Blackpearl's own purpose-built system led its results.
- **Non-commercial data in a commercial product.** HA-ViD's CC BY-NC licence and Te Hiku's Kaitiakitanga License both rule out unapproved commercial use.
- **Preprints read as settled.** The 2020 [temporal probability calibration preprint](https://arxiv.org/abs/2002.02644) co-written by Levy Street's Max Polaczuk has been cited by Google Brain researchers, but it is still not peer reviewed.

## Licences, Māori data and funding in New Zealand

Te Mana Raraunga's [Māori Data Sovereignty principles](https://www.temanararaunga.maori.nz/s/TMR-Maori-Data-Sovereignty-Principles-Oct-2018.pdf) say Māori data should be stored in Aotearoa whenever possible and that free, prior and informed consent should underpin its use. Te Hiku's licence applies that view to code and models.

For funding, Inland Revenue lists the [R&D Tax Incentive](https://www.ird.govt.nz/research-and-development/tax-incentive) as a 15% tax credit on eligible R&D, with a minimum of $50,000 of eligible spending a year. The Government committed up to $70 million over seven years to an AI research platform through the New Zealand Institute for Advanced Technology; as at 20 October 2026 we could not find an announcement naming the selected platform. If you want research turned into a working system, start with a [workflow assessment](/contact/?topic=assessment).

## Questions people ask

### Which New Zealand teams publish AI benchmarks or open source AI research?

Levy Street in Wellington publishes open-source code such as World of ClaudeCraft, an MIT-licensed game with an environment for training AI agents, and three of its team members co-authored GTM-Bench, a Blackpearl Group benchmark for sales-research agents. The University of Waikato maintains the open-source Weka and MOA tools, University of Auckland researchers published the LIBRA bias benchmark for New Zealand, and Te Hiku Media published a te reo Māori benchmark paper at NeurIPS 2024.

### Can I use Weka in a commercial project?

Weka is licensed under the GNU General Public License (GPL 3.0 for versions after 3.7.5). Waikato's own FAQ explains that any derivative work must also be licensed under the GPL if you distribute it to a third party. If you need to ship Weka code inside a product that cannot be GPL-licensed, it may be possible to buy a licence from the copyright holders, or to use the MIT-licensed tiny-weka subset. Get advice on your specific use.

### Are Te Hiku Media's te reo Māori speech models open source?

No. Te Hiku Media's code is under its Kaitiakitanga License, which says you must seek permission to access, use or modify the code and may not use it commercially unless Te Hiku explicitly grants that right. Its Papa Reo platform offers te reo Māori language technology through an API. Its NeurIPS 2024 benchmark paper states that it does not release new assets. Contact Te Hiku directly about access.

### Is there funding for AI research and development in New Zealand?

The main general scheme is the Research and Development Tax Incentive, which Inland Revenue lists as a 15% tax credit on eligible R&D, with minimum eligible spending of $50,000 a year. Separately, the Government committed up to $70 million over seven years to an AI research platform through the New Zealand Institute for Advanced Technology. As at 20 October 2026, we could not find an announcement naming the selected platform.

### Is GTM-Bench a Levy Street benchmark?

No. GTM-Bench is published by Blackpearl Group, and all eight authors are listed under Blackpearl Group Limited. Three of them, Max Polaczuk, Jeremy Baron and Reuben Horne, are on the Levy Street team. It has 72 tasks built from 59,881 real prospecting queries, and weak leads score negatively. The paper is self-published, not peer reviewed, and states that its production data environment limits reproducibility outside the benchmark operator.


## Sources

- [Te Ipu o te Mahara, Artificial Intelligence Institute](https://www.waikato.ac.nz/research/institutes-centres-entities/institutes/artificial-intelligence-institute/)
- [AI Institute software](https://www.waikato.ac.nz/research/institutes-centres-entities/institutes/artificial-intelligence-institute/research/software/)
- [AI Institute engagement](https://www.waikato.ac.nz/research/institutes-centres-entities/institutes/artificial-intelligence-institute/engagement/)
- [Weka](https://ml.cms.waikato.ac.nz/weka/)
- [Commercial applications](https://waikato.github.io/weka-wiki/faqs/commercial_applications/)
- [MOA repository](https://github.com/Waikato/moa)
- [CapyMOA](https://capymoa.org/)
- [Strong AI Lab](https://www.ai.ac.nz/)
- [Strong AI Lab repositories](https://github.com/Strong-AI-Lab)
- [Industrial AI Group](https://uoa-iai.github.io/)
- [HA-ViD: A Human Assembly Video Dataset](https://iai-hrc.github.io/ha-vid)
- [LIBRA: Measuring Bias of Large Language Model from a Local Context](https://arxiv.org/abs/2502.01679)
- [New Zealand AI research proposal reaches first stage](https://www.auckland.ac.nz/en/news/2025/12/18/new-zealand-ai-research-proposal-reaches-first-stage.html)
- [Centre for Data Science and Artificial Intelligence](https://www.wgtn.ac.nz/cdsai)
- [Evolutionary Computation and Machine Learning group](https://ecs.wgtn.ac.nz/Groups/ECRG/)
- [Spatial and Image Learning (SAIL) group](https://www.canterbury.ac.nz/research/about-uc-research/research-groups-and-centres/spatial-and-image-learning--sail--group)
- [Artificial Intelligence Research Group](https://www.canterbury.ac.nz/research/about-uc-research/research-groups-and-centres/artificial-intelligence-research-group)
- [Centre for Artificial Intelligence and Public Policy](https://www.otago.ac.nz/caipp)
- [Artificial intelligence research project](https://www.tepunahamatatini.ac.nz/our-research/artificial-intelligence)
- [Our research](https://www.tepunahamatatini.ac.nz/our-research)
- [Papa Reo](https://tehiku.nz/te-hiku-tech/papa-reo/)
- [Kaitiakitanga License](https://github.com/TeHikuMedia/Kaitiakitanga-License)
- [Fit for our purpose, not yours: Benchmark for a low-resource, Indigenous language](https://neurips.cc/virtual/2024/poster/97451)
- [New Zealand-led research at climate science's AI frontier](https://www.earthsciences.nz/news/new-zealand-led-research-at-climate-sciences-ai-frontier-unlocks-extreme-weather-insights)
- [About us](https://www.earthsciences.nz/about-us)
- [About AIRA](https://www.airesearchers.nz/about)
- [Aotearoa New Zealand Artificial Intelligence: A Strategic Approach, 3rd edition](https://www.airesearchers.nz/site_files/28243/upload_files/AIRA_Whitepaper_2026.pdf)
- [The NZ AI Benchmark](https://www.newzealand.ai/benchmark/)
- [Research and development tax incentive](https://www.ird.govt.nz/research-and-development/tax-incentive)
- [GTM-Bench paper](https://gtm-bench.ai/assets/gtm-bench-paper.pdf)
- [Temporal Probability Calibration](https://arxiv.org/abs/2002.02644)
- [World of ClaudeCraft repository](https://github.com/levy-street/world-of-claudecraft)
- [Principles of Māori Data Sovereignty](https://www.temanararaunga.maori.nz/s/TMR-Maori-Data-Sovereignty-Principles-Oct-2018.pdf)

Published by Levy Street (https://levystreet.com), Wellington, New Zealand.