Babbily adds Sakana Fugu Ultra, Kimi K2.7, and GLM 5.2

Babbily now supports Sakana Fugu Ultra, Kimi K2.7, and GLM 5.2, giving users access to three new advanced models built for reasoning, coding, long-context work, and complex AI workflows.

This release is part of our ongoing mission to make Babbily the easiest way to access the best AI models in one place. Instead of jumping between tools, subscriptions, interfaces, and providers, users can open Babbily and choose the right model for the job.

With this update, that choice gets even stronger.

Sakana Fugu Ultra: a different kind of model experience

Sakana Fugu Ultra is one of the most interesting new additions to Babbily because it does not behave like a typical single-model system.

Rather than relying on one model to solve every prompt, Fugu Ultra coordinates multiple frontier models. Depending on the complexity of the task, it can route work to different agents and then combine their outputs into one final answer.

That makes it especially useful for prompts where a second or third perspective can improve the result. For example, a user might ask it to review a pull request, analyze a technical paper, compare strategic options, or reason through a complex business decision.

In those cases, the value is not just raw speed. The value is stronger synthesis.

Fugu Ultra is a great fit for users who want deeper analysis, more thoughtful reasoning, and better handling of multi-step prompts.

Kimi K2.7: stronger coding and long-context development

Kimi K2.7 is another major addition for Babbily users, especially developers.

Kimi’s latest coding-focused model is designed for long-context programming tasks, stronger instruction following, and agentic workflows. That means it is a strong option for users who are asking AI to do more than generate small snippets.

It can help with debugging, reasoning through larger codebases, planning implementation steps, reviewing files, and completing complex programming tasks that require the model to stay focused over a longer session.

For Babbily users building apps, websites, automations, internal tools, or technical documentation, Kimi K2.7 gives them another strong model to reach for when coding quality and long-context reliability matter.

GLM 5.2: built for long-horizon engineering

GLM 5.2 is built for agentic software engineering and long-horizon technical work.

That makes it a strong fit for larger development tasks where the model needs to understand more context, keep track of more instructions, and reason across a longer workflow.

For users, this opens the door to more ambitious AI-assisted development inside Babbily. Instead of only using AI for quick answers or small code edits, users can use GLM 5.2 for bigger planning sessions, larger refactors, deeper technical reviews, and more structured engineering workflows.

It is especially useful when the task requires persistence, context, and step-by-step execution.

More models, more flexibility

The biggest benefit of this update is choice.

No single AI model is best at everything. Some models are better at creative writing. Some are better at coding. Some are better at long-context reasoning. Some are better at moving quickly. Others are better at deep analysis.

Babbily brings these models together so users can pick the right tool for the moment.

Need a multi-agent reasoning pass? Try Sakana Fugu Ultra.

Need stronger coding support? Try Kimi K2.7.

Need long-context engineering help? Try GLM 5.2.

This is what Babbily is built for: one AI workspace with access to the models that matter.

Available now

Sakana Fugu Ultra, Kimi K2.7, and GLM 5.2 are available now in Babbily.

Open Babbily, choose your model, and start building, writing, researching, coding, and creating with even more power.