Elan Mitsua is a Japanese machine learning-based art project with a V-Tuber mascot, producing image models (among others) trained only on appropriately licensed or explicitly submitted data.

Resolves YES if:

• I or any non-staff member is able to submit training data (image, text, or otherwise) for a future Mitsua model before 2026-01-09 0:00 JST (2026-01-08 15:00 UTC). The offset from the start of the year is to include a potential New Year's opening and delays.

Resolves NO if:

• Submissions are de facto disabled by the deadline, even if announced otherwise.

• The project is officially discontinued.

Abridged timeline so far:

• 2023, November 27 - Submissions for the next image generation model, Mitsua Likes, are opened.

• 2024, April 30 - Submissions are closed.

• 2024, November 26 - Beta versions of the model become open for use via Discord, for members who have applied to contribute.

• 2024, December 17 - Mitsua Likes is released publicly. A request form for the finetunable version (i.e. including VAE encoder weights) is opened, with applications "expected to be reviewed after 2025". Data submissions are said to reopen in January.

• 2025, February 10 - Staff apologizes for the delay in reopening submissions. They state that they were prioritizing training a Discord-exclusive version (on submitted data marked to be used for only that purpose; made available on February 4) and sending out digital rewards, and that they expected to complete both by the end of January.

• 2025, June 30 - Digital rewards for contributors are sent, with apologies for the delay. The rewards are: a thank-you video, a 3D model of the mascot, and the ability to apply for an "intensive study" model finetuned on the contributor's art, usable through Discord (separate from the VAE weights form; applications close July 18).

• 2025, August 15 - The "intensive study" finetunes are now available in the server. A second phase of applications is opened until August 29.