Back to list
Google Launches Gemini 3.5 Transcribe Featuring Automatic Filler Word Removal and Support for Over 85 Languages
Product LaunchGoogle GeminiAI TranscriptionAudio Intelligence

Google Launches Gemini 3.5 Transcribe Featuring Automatic Filler Word Removal and Support for Over 85 Languages

Google has officially expanded its Gemini Audio suite with the introduction of Gemini 3.5 Transcribe. This new AI-powered tool is designed to significantly enhance the quality of audio-to-text conversions by automatically filtering out common speech disfluencies, such as "ums" and "ahs." Beyond simply cleaning up speech, the model features advanced capabilities for detecting specialized technical jargon across various industries and offers robust support for more than 85 languages. This release follows the recent debut of Gemini 3.5 Live Translate, marking another step in Google's rollout of its 3.5-generation models. While the tech community continues to wait for the flagship Gemini 3.5 Pro model, Gemini 3.5 Transcribe provides a specialized solution for users seeking polished, professional, and linguistically diverse transcription services.

The Verge

Key Takeaways

  • Automated Speech Polishing: Gemini 3.5 Transcribe automatically detects and removes filler words like "ums" and "ahs" to create cleaner text outputs.
  • Extensive Language Support: The model supports transcription in more than 85 languages, catering to a global user base.
  • Jargon Detection: Advanced capabilities allow the AI to identify and correctly transcribe specialized industry-specific jargon.
  • Product Roadmap: This release follows Gemini 3.5 Live Translate, though the highly anticipated Gemini 3.5 Pro model is still pending release.

In-Depth Analysis

Refining Audio Intelligence with Gemini 3.5 Transcribe

Google's latest update to the Gemini Audio ecosystem, Gemini 3.5 Transcribe, represents a focused effort to move beyond literal transcription toward intelligent content refinement. One of the most prominent features of this new model is its ability to automatically edit out speech disfluencies. By identifying and removing "ums," "ahs," and other common filler words, the AI ensures that the resulting transcript is not just a record of what was said, but a professional document ready for immediate use. This functionality addresses a long-standing pain point in automated transcription, where raw text often requires extensive manual editing to remove the natural hesitations of human speech.

Furthermore, the model's ability to detect specialized jargon suggests a higher level of contextual awareness. In professional environments—ranging from medical and legal fields to high-tech engineering—standard transcription tools often struggle with niche terminology. By integrating jargon detection, Gemini 3.5 Transcribe aims to provide higher accuracy for specialized users, ensuring that technical terms are captured correctly without the need for constant manual correction.

Global Accessibility and Language Diversity

The scale of Gemini 3.5 Transcribe’s language support is a significant highlight of this release. With support for more than 85 languages, Google is positioning this tool as a comprehensive solution for international organizations and multilingual users. This broad support, combined with the model's ability to handle specialized terminology, indicates that Google is prioritizing the versatility of its AI tools across different cultural and professional contexts.

This release is strategically positioned within the Gemini 3.5 family. It follows the launch of Gemini 3.5 Live Translate, suggesting a modular approach to Google's AI rollout. By releasing specialized tools like Live Translate and Transcribe, Google is providing immediate utility to users in specific niches while the broader, more powerful Gemini 3.5 Pro model remains in development. This phased rollout allows Google to refine specific audio and linguistic capabilities before integrating them into a more comprehensive flagship model.

Industry Impact

The introduction of Gemini 3.5 Transcribe has several implications for the AI and transcription industries. First, it raises the bar for "out-of-the-box" transcript quality. As AI models become more adept at editing speech for clarity rather than just transcribing it verbatim, the demand for manual transcription and editing services may decrease. For businesses, this translates to faster turnaround times for meeting notes, interviews, and content creation.

Second, the focus on jargon and multilingual support highlights the growing importance of specialized AI. General-purpose models are increasingly being supplemented or replaced by tools that understand the specific nuances of different languages and professional sectors. Google's move to include these features in the Gemini 3.5 Transcribe model underscores a trend toward AI that is not only conversational but also technically precise and globally accessible. As the industry awaits Gemini 3.5 Pro, these incremental updates to the Gemini Audio suite demonstrate Google's commitment to maintaining a competitive edge in the rapidly evolving landscape of audio intelligence.

Frequently Asked Questions

Question: What are the main features of Gemini 3.5 Transcribe?

Gemini 3.5 Transcribe is designed to automatically remove filler words such as "ums" and "ahs" from audio transcriptions. It also features the ability to detect specialized industry jargon and supports more than 85 different languages, providing a more polished and accurate text output compared to standard transcription tools.

Question: How does this model fit into the Gemini 3.5 product lineup?

Gemini 3.5 Transcribe is a new addition to the Gemini family, following the launch of Gemini 3.5 Live Translate. It is part of Google's updated Gemini Audio capabilities. While these specialized tools are now available, Google has not yet released the Gemini 3.5 Pro model.

Question: Can Gemini 3.5 Transcribe handle technical or professional terminology?

Yes, one of the key features of Gemini 3.5 Transcribe is its ability to automatically detect and accurately transcribe specialized jargon, making it suitable for use in professional and technical fields where specific terminology is common.

Related News

Gemini API Tool Retrieval Preview Gives Agents Access to Hundreds of Tools With Fewer Tokens
Product Launch

Gemini API Tool Retrieval Preview Gives Agents Access to Hundreds of Tools With Fewer Tokens

Google AI Blog announced that Tool Retrieval is now in preview in the Gemini API. The feature is presented as a way to give agents access to hundreds of tools while using fewer tokens. The announcement pairs those two points: the breadth of tool access and lower token use. It does not explain how Tool Retrieval works, specify how token use was measured, or quantify the reduction. It also does not provide an exact tool count beyond describing access to hundreds of tools. The stated release status is preview, and the announcement does not give a timeline for general availability.

LangSmith Tool Gateway Enters Public Beta for Secure, Governed AI Agent Tool Use
Product Launch

LangSmith Tool Gateway Enters Public Beta for Secure, Governed AI Agent Tool Use

LangChain announced that LangSmith Tool Gateway is in Public Beta. The gateway handles credentials, identity, and access for every agent tool call, with policies, approvals, and traces. The announcement frames secure, governed tool use around the individual call: its focus is on how agents access tools, rather than on what those tools or agents can do. The update identifies what the gateway covers but does not explain how credentials are managed, how access decisions are made, when approvals apply, or what information traces contain. The Public Beta is the only release status stated in the announcement.

Google’s Gemini Decision 1 Picks Answers Instead of Writing Them Across Five Input Types
Product Launch

Google’s Gemini Decision 1 Picks Answers Instead of Writing Them Across Five Input Types

Google’s Gemini Decision 1 is a model designed to pick answers rather than write them, according to Shittu Olumide’s report for KDnuggets. It reads text, images, video, audio, and PDFs in a single forward pass and supports up to 262,144 input tokens per request. The central distinction is between what the model reads and what it does with that input: it accepts several kinds of material, but its stated task is answer selection, not answer writing. The token figure describes the maximum input per request, not how much the model outputs. The report does not explain how answers are presented for selection or provide results showing how well the model performs.