
What happened
GAZAI will begin providing an entertainment-focused LLM on 2026年11月11日 to game companies, VTuber agencies and publishers, drawing on its anini app and in-house GPU servers.
Why it matters
The company says the model generated natural Japanese character dialogue that scored equal to or better than an open model and a commercial model in five blind-rated criteria, so entertainment firms may get a ready-made option instead of building from scratch.
What to watch
The service starts on 2026年11月11日, with pricing available by individual quote.
WHO IT HITSEntertainment companies that own characters and worlds — game studios, VTuber agencies, manga app operators and publishers — may find a lower-cost route to add AI chat without building a model from zero, though pricing is quote-based and adoption is not yet proven.
Summaries like this, every morning.
GAZAI's move into enterprise LLM services follows the rollout of its consumer app anini, which began a closed beta in October 2025 and was officially released in April 2026; the app has since grown its user base and holds a 4.4 user rating on the App Store as of November 2026. The company is positioning the new service as a way to reuse the operational knowledge and technical base from that app for corporate clients.
The company says it has been accepted into the NVIDIA Inception Program and has built its own GPU server environment, which it says reduces reliance on commercial APIs and helps compress inference and training costs through specialized fine-tuning. That in-house setup underpins the service's promise of one-stop support, from model tuning on a client's own data through implementation and ongoing operational consulting.
In a blind evaluation of Japanese character dialogue, GAZAI's model was compared with an open model and a commercial API model. GAZAI says its model scored equal to or better than both across five criteria, including spatial and physical awareness, long-term conversation memory, descriptive balance of situation and action, five-sense depiction, and character-specific emotion and phrasing. The blind-test result is the company's own claim and has not been independently verified, so entertainment firms weighing adoption may want to run their own trials before committing.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
In Anthropic's IPO roadshow video, co-founder Tom Brown said Claude trains and serves across AWS Trainium, Goo…

Anthropic set a proposed IPO valuation of up to roughly $1.4 trillion, below earlier reported expectations of…

Databricks said Genie One now combines Genie Search across tools like Google Drive, Jira and Confluence, in-co…

Databricks introduced the new Genie, an autonomous AI coworker that stays on the job continuously, tracking pr…

The New York Times reported Jensen Huang allegedly had an affair with Alice Huang, a one-time Megaspeed execut…

Google DeepMind unveiled AlphaGenome 2, trained on more than 100 million individual cell profiles across 68 ti…
