Auto picks the model for each message

Choose Auto at the top of the model picker and stop wondering which model to use. For every message, Auto judges what kind of task it is and how demanding it is, then picks a model and reasoning level to match. A quick question can get a quick answer, and a hard problem can get a stronger model with more reasoning.

• New chats start on Auto when it is available for your account. Existing chats keep the model they were using.

• See which models answered. Auto · Models used: under the composer lists every model that replied in the chat.

• Choosing the model is free. It does not count toward your API usage budget, and when your budget runs low, Auto favors models that use less of it.

• Named models are one click away. Pick one whenever you want the same model every time, or want to set Fast mode or reasoning effort yourself. Those controls are hidden while Auto is selected.

Auto is rolling out to accounts gradually on the web. It is not offered in Translate or Scheduled Tasks. Learn how Auto works.

A cleaner, calmer design

Babbily has a quieter, more consistent look across chat, the sidebar, Settings, Library, Projects, News, Finance, and sign-in.

• A new typeface. Babbily now uses Geist throughout the app, and PDF and Word exports match.

• Stop lives in the composer. While Babbily writes, the send button becomes a stop button. Press Esc to stop, too. Stop a reply.

• A tidier thread. Your messages sit in soft rounded bubbles, replies start closer to your question, and Copy response, Good response, Bad response, Save to library, and Retry are compact icon buttons that stay out of the way on older replies.

• One header across every page. The title and Share line up the same way everywhere. In the sidebar, thread actions now live in a … menu, or right-click a chat. On a touch screen, press and hold.

• Built for every screen. Tablets get an always-visible icon rail, phones get a sliding drawer and bottom-sheet menus, and touch targets are larger.

• More accessible, and in more languages. A clearer focus ring, higher-contrast text, reduced-motion support, and pause buttons for the News and Finance tickers. Much more of Babbily now appears in all 10 app languages. Review accessibility support.

• A quieter notification offer. It is now a small card with Turn on notifications that you can close. Manage notifications.

New models: Claude Sonnet 5.5 and GPT-6.1 Sol

Claude Sonnet 5.5 is built for focused coding tasks, polished documents, and clear everyday work, with reasoning effort through Max. It is also available in Translate. GPT-6.1 Sol replaces GPT-6 Sol for demanding coding, reasoning, and complex professional workflows. It supports Fast mode and reasoning effort through Max. Compare current models.

A simpler model list

The picker now has 14 named models. Older models, including every model that was marked Sunsetting, have been removed.

• Images and videos come from built-in tools. Image and video models are no longer separate picker choices. GPT Image 2.5 Flare and Grok Imagine Video now power the Image generation and Video generation tools, which you can use from any chat model to create new media, edit an image, animate an image, or edit a video. Create images and videos.

• Removed models switch automatically. If a chat used a removed model, Babbily selects Auto or another available model and tells you. Nothing is resent; choose Retry if your last message did not get a reply. See removed models and alternatives.

• Translate has a new lineup. Choose Claude Sonnet 5.5, GPT-6 Luna (the new default), or Gemini 3.8 Flash. Translate uses each model’s lightest reasoning setting so results start quickly. Choose a translation model.

Also in this release

• Reasoning off uses the lightest setting. When reasoning is off, Babbily now asks each model for its lightest setting. Some models stop reasoning entirely; others, such as Claude Opus 5.5, still reason briefly at their lowest level. Gemini 3.8 Flash adds Low and High reasoning effort. Choose reasoning effort.

• Fewer unexpected sign-outs. A dropped connection or a temporary service problem no longer signs you out. If your account details do not load, Babbily shows We couldn’t load your account with Retry and Reload, and you stay signed in. Fix sign-in issues.

• Safer team invitations. If you open an invite link while signed out, Babbily asks you to confirm the email address the invitation was sent to. Team invitations no longer expire. Accept a team invite.

• Plugin approvals are clearer. Approval cards appear right away, typing “yes” in chat does not approve an action, and unanswered cards expire after 24 hours. Approve Plugin actions.

• Attach files in News and Finance. Add files to your question from an article or a Finance page.

• Memory understands your language. The memory box in Settings > Memory understands all 10 app languages. Manage memory.

• Leaner error reports. The reports we use to fix problems are now filtered to remove sign-in details, the messages you send, AI prompts and responses, and chat titles. Filtering is not a guarantee, so keep secrets out of chats and titles. See what error reports include.

Four new chat models are available in the model picker. Claude Opus 5.5 is suited to sustained coding and knowledge work. GPT-6 Sol handles demanding reasoning and coding, while GPT-6 Luna is a lighter option for focused work. Grok 4.7 supports coding and long-context research.

Choose Fast mode when speed matters. Claude Opus 5.5, GPT-6 Sol, and GPT-6 Luna have a Fast mode control. The selected model stays visible, and Fast mode can use more of your API usage budget.

Check older favorites before starting new work. Claude Opus 5, GPT-5.6 Sol, Terra, and Luna, and Grok 4.6 now show Sunsetting in the picker. They remain selectable during the transition. See current alternatives.

Use GPT-6 Sol in Translate. It replaces GPT-5.6 Sol in the Translate picker and becomes the default when your previous selection is unavailable. Claude Haiku 4.5 and Gemini 3.8 Flash remain available there.

Attachments no longer depend on the model you picked. Every chat model now accepts documents, text and code, images, audio, video, and ZIP archives. When the selected model cannot read a file itself, Babbily reads the file first and gives the model what it found, so you do not have to switch models before attaching something or start over because a thread has the “wrong” model for a file. See what you can attach.

More formats are supported. Excel (.xlsx) and PowerPoint (.pptx) files, RTF, HWPX, HEIC and HEIF photos from an iPhone, and a wider range of audio and video formats now work alongside PDF, Word, text, code, and images

ZIP archives work as one attachment. Attach a ZIP and Babbily reads the supported files inside it, up to 10 files per archive, and tells you which ones it skipped. Encrypted archives still need their password removed first. Attach a ZIP archive.

Each file is read once. Follow-up questions, message edits, and retries reuse what Babbily already read instead of reading the file again, so they are faster and do not spend your API usage budget on the same file twice. How reading uses your budget.

You can see what Babbily is doing. While a file is being read, the activity line shows Reading with the file name instead of a bare spinner.

Clearer answers when a file cannot be used. Babbily now names the file and the reason before the message is sent, and points to the fix: save a legacy .doc as .docx, remove a ZIP password, or send a smaller file. An old attachment that can no longer be read no longer blocks the rest of the thread. Troubleshoot a rejected file.

Long files fit better. File content now counts toward the context indicator as the text the model actually receives. When several files would not fit, Babbily shows the most recent ones in full and shortens the rest rather than failing your message. Understand the context window.

Chat replies now stream into the thread as Babbily produces them. Before this, a thread filled in steps while it waited for the next update, so the opening words took longer to land.

In our testing across a batch of more than 30 responses, replies began arriving about 47% sooner. How much faster your own chats feel depends on the model, the prompt, the tools a reply uses, and your connection.

Interrupted delivery recovers on its own. If the stream stalls or your connection drops, Babbily reconnects to the reply already in progress rather than starting a second one, so you do not get a duplicate answer.

Stop keeps what you already have. Choose Stop and the partial reply stays in the thread, including the text, reasoning, citations, and media produced so far. Stop also works in the moment right after you send, before any text appears. Stop a reply.

A stopped or stale thread cannot disturb a newer one. Stopping a reply, or a late update from a thread you have moved on from, no longer affects the conversation you are in now.

Now documented: replies keep running when you leave. Babbily has always generated replies on its servers rather than in your browser tab, so closing the tab does not cancel them. That behavior is unchanged, and it is now written down. Leave a reply running.

Keep related work together in Projects.

Create a Project in your Personal or team workspace to keep its chats, instructions, knowledge files, and project memory together. Project instructions shape every new chat, and reusable knowledge can include supported PDF, DOCX, text, Markdown, CSV, and JSON files. Team members can collaborate on shared project content while each chat remains continuable only by its author.

Run recurring prompts with Scheduled Tasks.

Paid plans can use Scheduled for manual, hourly, daily, weekday, weekly, or monthly prompts. Choose a model, time zone, and any connected Plugins the task may read from. Each run uses the API usage budget of its workspace and leaves its answer in a new chat under Recent runs.

Connect apps through Plugins.

Connectors is now Plugins throughout Babbily. The supported Plugin catalog now uses Composio for authorization, while Babbily continues to ask for approval before attended write actions run. Availability can vary by your app permissions, account state, and service status.

Saved Prompts is no longer available in Babbily. We made this change after seeing low usage and as Babbily continues to evolve toward more agentic ways of working. What this means:

Saved Prompts is no longer available.

You can no longer create, edit, or open Saved Prompts in Babbily.

Use memory for context you want reused.

Memory can carry stable preferences, work context, recurring projects, tone, audience, and formatting choices into future chats. This reduces how much setup you need to repeat.

Use prompt patterns for repeatable tasks.

Start with a consistent goal, context, audience, constraints, and output format. Keep task-specific instructions in the chat so you can adjust them as the work changes. You can also use a Skill when one fits your workflow.

Keep Finance context through follow-ups.

Start from a market overview, company, earnings event, or prediction market and continue asking questions without reselecting Finance research. Supported Finance threads preserve the page context and your tool choice across follow-ups, page reloads, and Retry.
​
​Control Finance research with Auto, Manual, or Off.

In a supported Finance thread, Auto uses Finance research automatically. Manual starts with Finance research enabled for that thread, and you can remove it. Off keeps the reply model-only.

Get stronger source coverage for financial questions.

Finance research now handles live quotes, historical periods, earnings, peer comparisons, and multi-company filing research more consistently. It checks requested coverage, keeps citations with the answer they support, and uses primary filings when available.
​
​See clear limits instead of unsupported figures.

When a company does not report a requested metric separately, Babbily explains what is unavailable and can suggest a related figure that is reported. Ambiguous company names or missing evidence may produce a clarification or limitation instead of a guess.
​
​Build full reports with the Deep Research Agent.

Choose Deep research for a broad question. Babbily plans sub-questions, opens the outline in Canvas, gathers and verifies sources, and streams a cited report while it works. Chat keeps a short summary, and the full report stays in Canvas for review, editing, and export.
​
​Keep useful work when a long run is interrupted.

If research reaches its time limit, Babbily keeps the report and tells you its coverage is partial. If report writing reaches its limit, the report is preserved and labeled incomplete. If a selected connector action needs approval, the request appears inline in the conversation; the report pauses and can continue after you approve it.

Babbily now keeps personal and team work in distinct workspaces. You can switch between Personal and every team workspace you belong to, collaborate with teammates without making work public, and keep an individual plan alongside a team plan.

What’s new

Switch between workspaces.

Open your account menu to choose Personal or a team workspace. The workspace pill in the sidebar shows where you are working. Chats, Saved Prompts, Library items, and memory stay with the workspace where you created them, so switching workspaces does not move existing work.

Share work with your team.

From a team workspace, you can share a chat or supported saved media with specific teammates or everyone in the team. Recipients find it under Shared with you in Library. Shared team chats are read-only, and access lasts only while the recipient remains a team member.

Use Shared team memory.

A team owner can turn Shared team memory on from Settings > Team. When it is on, useful facts from team-workspace chats can become shared context that current members can use in team chats. Personal memory remains separate. Authors and the team owner can delete eligible shared entries, and turning the setting off stops new additions.

Keep personal and team plans side by side.

Joining or creating a team no longer requires your individual plan to disappear. Your Personal workspace and team workspace show their own plan context, and owners manage a team plan from the team workspace they own. If a team plan becomes inactive, Babbily labels that workspace Plan inactive — read-only until access is restored.

Kimi K3 and Gemini 3.6 Flash.

Kimi K3 is available for long-horizon coding, agentic workflows, and large-context knowledge work. Gemini 3.6 Flash is available for fast coding, agentic work, and web development.

Three new frontier models are live in Babbily. Coding benchmarks below are from Cursor's own eval suite (cursor.com/evals), current as of today.

GPT-5.6 (Sol, Terra, Luna) — GPT-5.6 Sol Max ranks #3 overall at 67.2%, trailing only Claude Fable 5 Max (70.5%) and Fable 5 Extra High (68.4%). Terra Max scores 64.9% at roughly half the cost of Sol Max. Luna is the fast, low-cost tier.

Grok 4.5 — Grok 4.5 High scores 66.7%, ranking #4 overall, half a point behind GPT-5.6 Sol Max, at $1.51 per task versus $5.22, about 3.5x cheaper for nearly the same score.

Note: Cursor marks Grok 4.5's entries with an asterisk, tied to a reported training contamination issue (a Cursor codebase snapshot was inadvertently included in its training data). Worth checking Cursor's footnote before citing this externally without caveat.

Muse Spark 1.1 — Meta's upgraded flagship model, built for coding, debugging, multi-step agentic tasks, and multimodal reasoning across text, image, and video.

All three are available now in your model picker.