AutoScientist ChallengeOfficial Terms & Conditions

Organiser

The AutoScientist Challenge ("the Competition") is organised by Adaption, Inc. ("Adaption Labs"), operating under the marketing name Adaption ("Adaption", "we", "us", or "our"), 2261 Market Street STE 46020, San Francisco, CA 94114, USA.

Eligibility

2.1 The Competition is open to individuals and teams worldwide, subject to the exclusions in clause 2.2.

2.2 The following are not eligible to participate:

  • Employees, contractors, officers, or directors of Adaption, Inc. and their immediate family members
  • Residents of the province of Quebec, Canada
  • Residents of any jurisdiction where the Competition is prohibited by law
  • Individuals under the age of 18 (or the age of majority in their jurisdiction)

2.3 Teams may be of any size. Each individual may only participate in one team.

2.4 Participation is subject to acceptance by Adaption. Submitting an application does not guarantee entry. Accepted participants will be contacted directly.

Competition Period

3.1 Sign ups open on Monday, June 8, 2026 at 7:00 AM UTC.

3.2 The Competition runs in two consecutive parts:

  • Part 1 opens June 8, 2026 and closes June 22, 2026 at 11:59 PM UTC
  • Part 2 opens June 23, 2026 and closes July 6, 2026 at 11:59 PM UTC

3.3 Winners will be announced on July 10, 2026.

3.4 Adaption reserves the right to extend, shorten, or cancel the Competition Period with reasonable notice.

How to Enter

4.1 To enter, participants must:

  • Submit an application on the sign-up page and receive confirmation of acceptance
  • Select one of the 10 competition categories and indicate whether they are entering Part 1 or Part 2
  • Train and submit an original frontier model during the relevant Competition Period
  • Publish both the adapted fine-tuning dataset and the trained model weights publicly to Hugging Face and Kaggle
  • Credit AutoScientist by Adaption in the model and dataset cards

4.2 No purchase is necessary to enter or win.

Judging

5.1 Submissions will be evaluated by a panel selected by Adaption. The primary judging criterion is measurable percentage improvement against a baseline model on Adaption's held-out in-house test set for each category.

5.2 Each part uses a separate held-out test set covering its five categories. These test sets are not disclosed to participants at any time.

5.3 Bonus consideration will be given to participants who:

  • Build and publicly release a demo of AutoScientist
  • Share their project on LinkedIn and X tagging @adaption_ai and adaption-labs

5.4 Judges' decisions are final and binding. No correspondence will be entered into regarding judging outcomes.

5.5 Adaption reserves the right to withhold prizes if submissions do not meet a minimum quality threshold.

Credits

6.1 Accepted participants will receive 1,000 Adaptive Data credits for AutoScientist to use throughout the Competition Period.

6.2 Credits are non-transferable and have no cash value.

Prizes

7.1 The total prize pool is USD $50,000, distributed as follows:

First Prize: USD $4,000 x 10 (one per category)

Runner Up: USD $1,000 x 10 (one per category)

Honorary Mention: credits and swag x 30

7.2 Prizes are awarded per category across both Part 1 and Part 2 combined. Winners are announced on July 10, 2026.

7.3 Prizes will be paid via wire transfer within 20 days of winner verification.

7.4 Prizes are non-transferable and cannot be exchanged for cash alternatives, except at Adaption's sole discretion.

Intellectual Property

8.1 All submissions must be open-sourced under a licence approved by Adaption (e.g. CC BY 4.0 or Apache 2.0) and published to both Hugging Face and Kaggle.

8.2 This includes both the adapted fine-tuning dataset and the trained model weights.

Participant Warranties

By entering, each participant warrants that:

  • They are eligible to participate under these Terms
  • Their submission is original and does not infringe any third-party rights
  • Their submission does not contain unlawful, harmful, or offensive content
  • They have all necessary rights to submit and open-source both the dataset and model weights

Disqualification

Adaption reserves the right to disqualify any participant who:

  • Breaches these Terms
  • Engages in fraudulent, misleading, or unsportsmanlike conduct
  • Attempts to manipulate or game the judging process or held-out test sets
  • Submits content that infringes third-party rights

Limitation of Liability

11.1 To the fullest extent permitted by law, Adaption accepts no liability for any loss, damage, or injury arising from participation in the Competition.

11.2 Adaption is not responsible for technical failures, late or lost submissions, or circumstances outside its reasonable control.

Data Protection

Adaption will process participants' personal data in accordance with applicable data protection laws including GDPR, CCPA, and PIPEDA where applicable.

General

13.1 These Terms are governed by the laws of the State of California, USA, without regard to conflict of law principles. Any disputes arising out of or in connection with these Terms shall be subject to the exclusive jurisdiction of the courts of San Francisco County, California.

13.2 Adaption reserves the right to amend these Terms at any time with reasonable notice.

13.3 If any provision of these Terms is found to be unenforceable, the remaining provisions shall continue in full force.

13.4 These Terms constitute the entire agreement between participants and Adaption in relation to the Competition.

Contact

For questions regarding these Terms, contact team@adaptionlabs.ai.

Last updated: June 4, 2026