Find the best tweets about Google Gemini, covering new models, multimodal features, developer tools, benchmarks, and practical workflows. Updated every two weeks.
Hands-on Gemini use, product releases, research, and comparisons backed by specific examples.
Original Xholic analysis
Across 50 tweets from 39 creators, Gemini conversation skews strongly favorable: 88% of posts are supportive (44 of 50) and 68% carry positive sentiment, with 24% neutral and 8% negative. Announcement-format posts make up 68% of the conversation. The standout post is Shushant Lakhyani's hands-on Gemini headshot-generation tutorial (2036820521371967624), which posted an all-time score of 7057.7 β roughly 338x the community median of 20.86. The strongest-performing themes by median all-time score are agentic capabilities & computer use (112.11), Gemini API capabilities & developer integrations (101.7), and model releases, benchmarks & pricing comparisons (65.73). A caveat for a hands-on angle:
68% of posts
All-time engagement
68% of posts
Published in 90 days
Conversation map
Step-by-step user examples of getting work done with Gemini: corporate headshot image generation, VFX with Omni/Flow, Deep Research reports, NotebookLM integration, Gemini in Chrome, customer sentiment analysis, and investor pitch stressβt
20%
Launches and upgrades to Gemini Live and Flash Live audio models: latency, interruption handling, tone/emotion understanding, live translation, Search Live expansion, and interactive news briefs.
18%
Announcements and third-party evaluations of new Gemini model versions (3.1 Pro, Flash-Lite, 3.5 Flash, 4 Argon), including benchmark scores, cost-per-task, pricing tiers, and comparisons against GPT/Claude/Grok models.
16%
User assessments of Gemini for coding: complaints about Gemini CLI reliability, quota limits and AI Ultra disappointment, plus reports on Google's push to catch up with Codex and Claude Code.
12%
New Gemini API features such as built-in tools (Search, Maps, File Search, URL Context) combined with function calling, the gemini-api CLI, skills for up-to-date SDK knowledge, migration guides, and AI Studio/Vertex availability.
12%
Weekly Gemini CLI release notes, extensions, subagents, plan mode, and hands-on setup/usage of the terminal agent, including integration with MCP servers and Google Workspace.
12%
Curated lists and courses for learning Gemini: Gemini Academy, API docs, cookbooks, prompt guides, free vibe-coding course, and Google's open-sourced agent skills library usable across coding tools.
6%
Gemini acting autonomously on behalf of users: Gemini Desktop full-access/sandbox permissions, agent benchmarks (AutomationBench, Terminal Bench), and agent-building with ADK.
4%
Tone and stance
Performance benchmark
Posts with media make up 74% of this collection. Their median all-time score is 32.3, compared with 2.72 for text-only posts.
Format mix
Consensus and debate
Shared view
Multiple independent posts cover the Gemini 3.1 Flash Live launch β lower latency, better tone/emotion understanding, native interruption handling, and 2x longer conversational context β powering both Gemini Live and the global expansion of Search Live.
Shared view
Posts converge on newly shipped API capabilities: built-in tools (Search, Maps, File Search, URL Context) combined with function calling in a single call, the gemini-api CLI for image/video/TTS/music generation and media analysis, and Gemini API skills that lifted Gemini 3.1 Pro to 95β96.6% success across 117 coding eval prompts.
Shared view
Announced with 77.1% on ARC-AGI-2, described as more than 2x Gemini 3 Pro, rolling out across the Gemini API, AI Studio, Vertex AI, Gemini Enterprise, the Gemini app, and NotebookLM.
Shared view
Tutorials and recipes cover corporate headshot generation, VFX via Omni Flash in Google Flow, Deep Research inside NotebookLM (with source separation and report export options), Gemini in Chrome, NotebookLM-to-Gemini writing workflows, and using Gemini to stress-test investor pitches as a skeptical VC.
Open debate
BridgeMind calls Gemini CLI with Gemini 3.1 Pro 'one of the worst vibe coding tools on the market' β citing mid-session crashes, lost context, unreliable multi-file tasks, and cancelling a $250/month AI Ultra plan β while separately noting the model itself is a solid baseline that ranks #1 on BridgeBench Hallucination. Andrew Brown reports similarly erratic results on a simple hello-world install test.
Open debate
Alex Heath reports Google faces internal pressure to catch up on coding, with a then-upcoming model in the same ballpark as GPT-5.5; Mark Kretschmann relays Logan Kilpatrick's statement that Google will stop letting Codex and Claude Code 'eat their lunch,' while observing Gemini CLI is usable but not very popular.
Open debate
AshutoshShrivastava argues Gemini's ~$20 tier (latest model access, NotebookLM, Antigravity, 5TB storage, Gemini CLI/Jules/Code Assist, Workspace integration, 1,000 monthly AI credits) is the best subscription available, while another user reports AI Ultra capacity limits and reduced quotas that drove them toward cancellation.
Open debate
Artificial Analysis finds Gemini 4 Argon matches GPT-6 Astra on the Intelligence Index at 60% of the cost per task, ranks #1 on AutomationBench-AA (77.5%), and posts the lowest hallucination rate (15%) among leading models β but is not yet publicly available and reverts to $3.98/task when the 50% launch discount ends. In contrast, a critical review argues Flash underperforms Grok on agentic loops and that Google 'l'
What performs
Median all-time score for posts with media is 32.281 vs 2.719 for text-only β roughly a 12x gap β and 74% of posts include media.
ANNOUNCEMENT posts (68% share) have a median all-time score of 48.87, versus TUTORIAL at 1.24 and LIST at 4.078 β though the single largest outlier is a tutorial.
The five highest-scoring posts: the headshot tutorial (7057.7, 338x median), Sundar Pichai's Gemini 3.1 Pro announcement (590.42, 28.3x), Artificial Analysis's Gemini 4 Argon evaluation (542.94, 26.03x), Logan Kilpatrick's Gemini 3.5 Flash Live Translate launch (474.31, 22.74x), and Dmitry Lyalin's gemini-api CLI announcement (472.07, 22.63x).
The lone QUESTION post β asking readers to name a better subscription than Gemini at $20 β scored 221.137, the highest median of any format, showing conversational prompts can rival announcements.
Median engagement per post: 94 likes, 6,453 views, 12 replies, 13 reposts, and an all-time score of 20.86 β meaning most Gemini content performs near baseline while a handful of launches and tutorials dominate totals.
Statistical standouts
Creator landscape
The five most represented creators account for 20% of the selected posts.
1. Philipp Schmid
@_philschmid
2 posts
2. Vaishnavi
@_vmlops
2 posts
3. AshutoshShrivastava
@ai_for_success
2 posts
4. Glenn Gabe
@glenngabe
2 posts
5. Julian Goldie SEO
@JulianGoldieSEO
2 posts
6. Mark Kretschmann
@mark_k
2 posts
The Google Gemini lead posted twice with a median all-time score of 289.36, including the Gemini 3.5 Flash Live Translate launch (474.31) and Gemini API built-in-tools update (104.4).
Two posts with a median all-time score of 118.08, including the subscription-value question (221.137) and a Gemini CLI subagents update (15.015).
Two posts with a median all-time score of 85.4, covering the Gemini API skills blog (99.006) and the built-in-tools-plus-function-calling launch (71.787).
Since the previous snapshot
Themes, sentiment, stance, and post format are classified per tweet. All counts, shares, medians, creator concentration, freshness, and performance comparisons are then calculated directly from the published snapshot.
Xholic's all-time score compares engagement while accounting for reach, post age, and creator consistency. It is used for relative comparisons within this collection.
This report analyzes the exact 50-post snapshot shown below. AI identifies editorial categories and drafts explanations; all statistics are calculated from the snapshot, and every narrative claim is checked against cited posts before publication.
Best Google Gemini tweets
Ranked 01β50
@shushant_l Β·
Here's how to create your classic corporate headshot: 1. Go to Google Gemini 2. Select 'Pro' 3. Select 'Create image' tool 4. Add your image 5. Add this prompt: cinematic, ultra-realistic corporate headshot of the same person from the reference image. Preserve exact facial features, skin tone, and identity. Style: classic professional corporate portrait. The subject is wearing a tailored dark suit, crisp white shirt, and subtle tie. Lighting is soft, directional studio lighting with gentle shadows for depth. Background is minimal, blurred, and neutral-toned (dark grey or beige gradient). Shot on a high-end full-frame camera with 85mm lens, shallow depth of field, sharp focus on the eyes, natural skin texture, and high dynamic range. Expression is confident, calm, and approachable. Ultra-detailed, 8K quality, realistic color grading 6. Submit 7. AI generates in seconds
@sundarpichai Β·
Gemini 3.1 Pro is here. Hitting 77.1% on ARC-AGI-2, itβs a step forward in core reasoning (more than 2x 3 Pro). With a more capable baseline, itβs great for super complex tasks like visualizing difficult concepts, synthesizing data into a single view, or bringing creative projects to life. Weβre shipping 3.1 Pro across our consumer and developer products to bring this underlying leap in intelligence to your everyday applications right away. Rolling out now to: - Developers in preview via the Gemini API in @GoogleAIStudio - Enterprises in Vertex AI and Gemini Enterprise - Everyone through the @Geminiapp and @NotebookLM
@LyalinDotCom Β·
Introducing Gemini API CLI (gemini-api), now available in preview! Our CLI gives your agent power to generate images, videos, text-to-speech, or music, analyze media, transcribe audio, or manage agents. No more guessing model strings, throwaway scripts or endless curl. 𧡠π
Watch video
@ArtificialAnlys Β·
Googleβs new Gemini 4 Argon equals GPT-6 Astra on the Artificial Analysis Intelligence Index at 60% of the Cost per Task with discounted prices. Google is now back to being one of the top three labs in intelligence achieved Gemini 4 Argon is @GoogleDeepMindβs first proprietary model above the Flash class in over 7 months. With high reasoning (the highest available), it scores 53 on the Artificial Analysis Intelligence Index, matching GPT-6 Astra (max, 53) and 1 point ahead of GPT-6.1 Sol (max, 52), with gains driven by lower hallucinations and stronger agentic capabilities. At its current 50% pricing discount and with cache discounts increased to 95%, Gemini 4 Argon costs $1.99 per Intelligence Index task, 60% of GPT-6 Astra (max), but 2.7x GPT-6.1 Sol (max). After the discount ends, this will rise to $3.98 (~1.2x GPT-6 Astra (max)). Gemini 4 Argon is currently being rolled out to selected users and is not publicly available. The 50% discount is an initial promotion. Google has not yet confirmed the promotion end date Key benchmarking results for Gemini 4 Argon with high reasoning: β€ Google returns as one of the top three labs on intelligence: Gemini 4 Argon (high) scores 53 on the Artificial Analysis Intelligence Index, matching GPT-6 Astra (max, 53) and 1 point ahead of GPT-6.1 Sol (max, 52). This is 23 points above Googleβs previous non-Flash model, Gemini 3.1 Pro Preview (30) and 12 points ahead of Gemini 3.8 Flash (high) β€ Launch discounts of 50% make Gemini 4 Argon competitive on Cost per Task: At current discounted pricing, Gemini 4 Argon (high) costs $1.99 per Intelligence Index task, 60% of GPT-6 Astra (max, $3.26) for a comparable level of intelligence. This cost efficiency is driven by lower token prices, rather than reduced token use, with Gemini 4 Argon averaging 62k output tokens per task, compared with 27k for GPT-6 Astra (max). Google has not yet confirmed the promotion end date, but on standard pricing, Cost per Task will increase to $3.98 β€ Stronger agentic performance: Historically a weaker area for Gemini models, Gemini 4 Argon shows improvements across agentic evaluations. It ranks #1 on AutomationBench-AA at 77.5%, 6 points ahead of Claude Sonnet 5.5 (max, 71.3%). On Terminal Bench 4, Gemini 4 Argon achieves 57%, a +53 point improvement from Gemini 3.1 Pro Preview, only behind Claude Sonnet 5.5 (max, 64%), Claude Opus 5.5 (max, 60%) and GPT-6 Astra (59%). On AA-Briefcase, it reaches 1494 Elo. This is driven by a 65% rubric pass rate, the highest we have recorded, but lower Analytical Quality (1576 Elo) and Presentation Quality (1308 Elo) β€ Lowest hallucination rate among leading models: On AA-Omniscience, Gemini 4 Argon has a 15% hallucination rate, the lowest of any model scoring 45+ on the Intelligence Index, compared with 51% for GPT-6 Astra (max) and 54% for GPT-6.1 Sol (max). This means Argon is much more likely to acknowledge when it does not know an answer rather than guess incorrectly. On accuracy, Gemini 4 Argon scores 50%, a 5 point decrease from Gemini 3.1 Pro Preview, and 13 points below GPT-6 Astra (max, 63%). With this slightly lower accuracy, its overall AA-Omniscience score of 42 remains in line with GPT-6 Astra (43) and GPT-6.1 Sol (42) Key model details: β€ Context Window: 1M tokens β€ Multimodality: Text, image, video, and speech input, with text output β€ Pricing: $4/$20 per 1M input/output tokens at standard pricing, currently discounted 50% to $2/$10. Cached input tokens receive a 95% discount ($0.10 per 1M at discounted pricing), up from 90% on Gemini 3.8 Flash β€ Long Decode Continuation: We tested Gemini 4 Argon with Long Decode Continuation, a new Gemini API feature that pauses long responses and resumes them across follow-up calls. This lets reasoning run up to 1M output tokens without request timeouts
@OfficialLoganK Β·
Introducing Gemini 3.5 Flash Live Translate, our real time speech to speech translation model which supports more than 70 languages (both in and out), and is so natural. It is available in the Gemini API, AI Studio, & Google Translate right now + coming soon to Google Meet!!
Watch video
@GeminiApp Β·
Gemini Live just got its biggest upgrade yet, powered by Gemini 3.1 Flash Live. β’Faster responses with fewer awkward pauses β’Smarter & able to follow along 2x longer conversations, so you can stay in the flow β’Dynamically adjusts its answer lengths & tone to match the moment
Watch video
@testingcatalog Β·
GOOGLE π₯: Gemini Desktop will have a "Full Access" permission for computer use, allowing it to access any file and run any app on your desktop. > A new "Additional sandbox options" setting has been added to the recent Gemini Desktop app (Still hidden). > "By enabling additional sandbox options, you will be able to expand what Gemini can do and access on your Mac. Depending on which settings you enable, Gemini may be permitted to take actions without asking for your permission first, including: - Reading, creating, modifying, or deleting files anywhere on your Mac, including files outside your connected folders and files belonging to other people stored on your device - Sending and receiving data over the network without your approval for each connection, including accessing websites, APIs, and services you are logged into - Communicating with other applications on your Mac (such as Mail, Safari, or Messages) and performing actions through them." > Gemini will still ask user permission before purchasing products, creating accounts, accepting legal terms, or modifying sensitive information about you. It is very likely that computer use features are expected to be powered by Gemini 4. Soon? π h/t @bedros_p
@ai_for_success Β·
Name me a better subscription than Gemini for 20 dollars. Now 5TB Storage from 2TB. You get: 1. Gemini latest model access 2. NotebookLM 3. Antigravity 4. 5 TB cloud storage 5. Gemini CLI, Jules and Code Assist 6. Gemini in Gmail, Docs, Vids, and more 7. 1,000 monthly AI credits. I must have definitely missed few things here..
@HeyZaraKhan Β·
Become Google AI Expert (for free). The people learning this now will have a massive advantage in the next 3 years. Here are all the best resources in one place: (save this) 1. Gemini Academy: https://t.co/Kg7C1aV7ZS 2. Gemini API Documentation: https://t.co/RJCyVN350Z 3. Prompt Engineering Guide: https://t.co/aw7QqgrhMf 4. Gemini Cookbook: https://t.co/qsfgyWY9yu 5. Google AI Studio: https://t.co/21h89yrpPT 6. Gemini for Developers YouTube: https://t.co/wblLKguBcg Vertex 7. AI Documentation: https://t.co/wqDHKQa4ZG 8. Gemini SDK Examples: https://t.co/jelMSMu39n 9. MCP Documentation: https://t.co/V07VmChEf6 10. Free AI Courses by Google: https://t.co/XBQWkQ4Ppt I hope you found this helpful. For more you can follow me @HeyZaraKhan
@joshwoodward Β·
New in Gemini: Live's biggest upgrade yet Faster responses. Smarter responses. More EQ. More linguistic range. 2x longer context. Android and iOS, powered by Gemini 3.1 Flash. Enjoy!
Watch video
@geminicli Β·
Gemini 3.1 Flash-Lite is our most cost-efficient Gemini 3 series model π¦ Folks who are keen to experiment with it in Gemini CLI can do so with an API key and the -m flag. β― gemini -m gemini-3.1-flash-lite-preview Official support for it in /model will be coming soon π
Watch video@sundarpichai Β·
Gemini 3.1 Flash Live is our highest-quality audio and voice model yet. Voice capabilities have come a long way and are a big part of how we interact with AI to get things done. 3.1 Flash Liveβs improved precision and reasoning make those interactions more natural and intuitive. Available in @GoogleAIStudio through the Gemini Live API in preview.
Watch video
@googleaidevs Β·
Give your coding agents up-to-date SDK knowledge with Gemini API skills. Our evaluation saw 96.6% success with the gemini-3.1-pro-preview model across 117 coding prompts, measuring model performance across agentic, chat, and SDK use cases.
@erinkwoo Β·
SCOOP: Google's Gemini model hacked three companies as part of a May cybersecurity evaluation conducted by the testing company Irregular. Google was notified about the hacks in July, but didn't disclose them until we reached out this week. w @bobmcmillan:https://t.co/ohjbpyfVGg
@SriThreePO Β·
Here is Gemini CLIβs January 19th weekly update for v0.25.0 - π½οΈGemini CLI x https://t.co/MTzjXcNvyL course: A free course to take you from installation to advanced use-cases! Code & Create with an Open-Source Agent: https://t.co/1xjj8STsiR - πGemini CLI Extensions: - Pickle Rick π₯: Turn Gemini CLI into a brilliant, arrogant engineer. This extension uses the Ralph Wiggum technique to automate the SDLC. gemini extensions install https://github.βcom/galz10/pickle-rick-extension https://t.co/V5z7UfmbPv - Visa : Empower AI agents to securely browse, shop, and execute purchases, enabling autonomous commerce experiences on behalf of end-users. gemini extensions install https://github.βcom/visa/mcp - πCLI Help Agent: Gemini CLI specific requests are handed off to a new help agent to provide information on what is possible and how to configure your experience. - Terminal Title Agent Status: Terminal title will now show β Ready, β¦ Workingβ¦ or β Action Required depending on agent state. - LLM Edit Correction Control: For faster edits opt-out of automatic LLM-based edit correction by setting tools.disableLLMCorrection to true. - Interactive Shell Hotkey Change: Swapped from ctrl + F to Tab to toggle focus between the input area and the interactive shell. - In CLI Extension Management: Install and uninstall extensions in the CLI via /extensions install|uninstall - Slash Command File Selection: Slash commands now support using @ file selector to reference and insert file paths. π§΅
@OfficialLoganK Β·
Lots of great Gemini API updates shipping today π οΈ 1. Built-in tools (search, maps, file search) now work with function calling 2. We now do context circulation with built-in tools for better model performance 3. Grounding with Google Maps now works with Gemini 3!!
@mark_k Β·
Google Gemini is about to make a big comeback in AI coding, according to Logan Kilpatrick from @GoogleDeepMind. They are no longer going to watch Codex and Claude Code eat their lunch. Google already has the "Gemini CLI" tool, which is usable but not very popular at this time.
@_philschmid Β·
We just published a blog on how we built the Gemini API skill. LLMs have fixed knowledge cutoffs, so we need to teach them about our newest models and how to use the SDK. In our evaluations, it helped Gemini 3.1 Pro pass 95% of 117 eval tests. Skills and Blog below
@_philschmid Β·
We shipped one of the most requested Gemini API features! π₯³ You can now combine built-in tools (Google Search, URL Context,β¦) with your own functions custom in a single API call. Gemini orchestrates everything: π§Β Combine Google Search, Google Maps, File Search or Url Context with custom tools in one API call π Built-in tool context is circulated with signatures. π€ Gemini decides tool order and chains results. π Google Maps is now available for Gemini 3 models. Available natively in the Interactions API and opt-in via generate_content. Learn More below π½
@bridgemindai Β·
Gemini CLI with Gemini 3.1 Pro is one of the worst vibe coding tools on the market. The model itself is a solid baseline. #1 on BridgeBench Hallucination. Decent coding intelligence. But the tooling around it is terrible. Crashes mid-session. Loses context. Can't complete multi-file tasks reliably. I cancelled my $250/month Google AI Ultra plan for a reason. Google has a competitive model wrapped in unusable infrastructure. Claude Code with Claude Opus 4.6 and Codex with GPT 5.4 are years ahead on developer experience. Who else is having problems with Gemini CLI?
@VaibhavSisinty Β·
Google just open-sourced 13 official skills for AI agents. And they work with Claude Code, Cursor, and Copilot. Not just Gemini. π€― Read that again. Google built skills that make their competitors' tools better. One command to install: npx skills add google/skills Pick the ones you need. They load automatically. What's inside: β BigQuery, Cloud Run, Firebase, AlloyDB, Cloud SQL, GKE basics β Google Cloud onboarding, authentication, network observability recipes β Security, reliability, cost optimization from the Well-Architected Framework β Gemini API and Managed Agents Platform skills These aren't docs dumped into a prompt. They're compact expertise packs that agents load on-demand. No context bloat. No hallucinating old SDKs. Google didn't lock this behind Gemini. They made it work everywhere. That's the most un-Google thing Google has done in a while. https://t.co/eVldum25Uz
@alexeheath Β·
Sources: Google plans to announce a new Gemini model next week. Itβll be in the same ballpark as GPT-5.5 in terms of performance. Iβm told that inside Google, thereβs real pressure to catch up on coding capabilities in particular. https://t.co/UQ8sI1zKu0
@JeffDean Β·
π’ Another exciting step forward today with the launch of Gemini 3.1 Flash Live. It natively understands audio, making it much more capable of handling complex instructions. It leads on ComplexFuncBench, and on Scale AIβs AudioMultiChallenge, demonstrating skill in complex instruction following and long-horizon reasoning amidst the interruptions of real-world audio. That means the model can pick up on nuances like pitch and pace, leading to much more fluid, high-fidelity voice interactions. Itβs now powering Gemini Live and Search Live globally. Huge congrats to the teams that made this possible. β¨ https://t.co/5ayNyrJak1
@jerrod_lew Β·
Google Gemini Omni is great for VFX. Film anything on your phone and upload it to Google Flow. Select Omni Flash and insert a prompt to direct what happens in the scene. Try to give as much specific details as you can. Have left the prompt used for this video in the comments!
Watch video
@_vmlops Β·
GOOGLE DROPPED AN AI AGENT THAT LIVES IN YOUR TERMINAL and it's completely free meet gemini-cli most AI tools make you leave your workflow to use them gemini-cli doesn't it lives where you work the terminal and it doesn't just chat...it acts here's what it does: β«οΈ reads + edits your entire codebase β«οΈ runs shell commands natively β«οΈ fetches web pages in real time β«οΈ grounds answers with google search β«οΈ generates apps from pdfs and images β«οΈ saves and resumes complex sessions β«οΈ plugs into CI/CD scripts non-interactively β«οΈ extends with MCP servers for custom tools the free tier alone is insane: β«οΈ 60 requests/min β«οΈ 1,000 requests/day β«οΈ 1M token context window β«οΈ no api key needed.... just your google account even your github workflow gets smarter: trigger it on PRs for automated code review no setup headaches open source...apache 2.0.. this is what an AI-native terminal experience looks like `npx @google/gemini-cli` π https://t.co/N1YzXAlHPB
@SriThreePO Β·
Here is Gemini CLIβs February 25th weekly update for v0.30.0 - β¨Gemini 3.1: - Over the past week we started rolling out Gemini 3.1. In the coming days, all usage of Gemini 3 will automatically shift to Gemini 3.1, ensuring you have access to our latest and most capable model by default - β¨Gemini 3 by Default: To make it more discoverable, we have removed the previewFeatures requirement for our latest models. Gemini CLI now defaults to Auto (Gemini 3) family and you can continue to choose Auto (Gemini 2.5) - For Gemini Code Assist users your administrator will still need to enable Preview Release Channels for you to access Gemini 3. - πGemini CLI Extensions: - Neo4J: Generate Cypher queries using natural language, and build sophisticated GraphRAG applications within Gemini CLI - gemini extensions install https://github.βcom/neo4j-contrib/mcp-neo4j - Atlassian: Connect Atlassian Rovo MCP Server with Gemini CLI to sync, search, and automate Atlassian apps such Jira, Compass, and Confluence using natural language - gemini extensions install - https://github.βcom/atlassian/atlassian-mcp-server - π±οΈ Enhanced Drag-and-Drop Workflow - Giving you more seamless file handling when you drag and drop. - Cross-Platform Reliability: Dropping files into the CLI is now smoother and more reliable than ever across macOS, Windows, and Linux. - External File Support: You can now drag and drop files from outside your current workspace. To keep things secure, the CLI will explicitly ask for permission before accessing files from external directories. - π€Inline Thinking: Thinking bubbles now support summary and full modes for better visibility into the model's process. - π€Ctrl+z to Suspend - Posix users can now suspend by hitting ctrl+z and take back their terminal. Once you're ready to resume, you can foreground it with the fg command. -β³Default Retention Rates - Chat sessions have a 30 day TTL by default now. A custom retention can be configured via /settings βKeep chat historyβ - π Policy Engine Transition: We are deprecating --allowed-tools and excludeTools in lieu of policy engine in the coming weeks. - πΊοΈ Plan Mode (Experimental): - π§ Sequential Planning: Formalized a new 5-phase sequential planning workflow to enhance agentic execution and strategy.
@_vmlops Β·
GOOGLE JUST DROPPED A FREE "VIBE CODING" COURSE FOR BEGINNERS no coding experience needed to build real apps with gemini β 6hrs 15min, self-paced, introductory level β learn to build software using google's gemini AI β earn a skill badge on completion β challenge lab option to skip straight to testing your skills if you're already confident anyone curious + motivated can go from zero to shipped app
@mark_k Β·
Google launched Gemini 3.1 Flash Live today, its highest-quality real-time audio model for natural voice conversations. The model powers enhanced Gemini Live experiences and is available in preview through the Live API for developers. Main new capabilities: - Significantly lower latency for smoother dialogue - Improved tone and emotion understanding - Strong performance in noisy environments - Real-time multimodal support for audio and vision - Native interruption handling and barge-in - Expanded language coverage and tool integration The update makes voice AI faster, more reliable, and more human-like for users worldwide.
@ai_for_success Β·
Gemini CLI Update : Google has released subagents for Gemini CLI, allowing the tool to delegate complex and high-volume tasks to specialized expert agents. You can create custom subagents using Markdown and YAML It also Includes built-in agents: generalist, cli_help, and codebase_investigator Video Credit : Google blog
Watch video
@aiedge_ Β·
The new Google Gemini skills are actually very valuable. I've been automating my Chrome browser tasks with stuff like: β’ Summary skills β’ Shopping/price skills β’ Travel/planning β’ Research/news sourcing Google also has a ready-to-use Skills library - highly recommend:
@brodieseo Β·
Bad idea to underestimate Google. Gemini's global market share just reached 21.5% as of today. 3 months ago: 12.9% 6 months ago: 8.6% 12 months ago: 5.7%
@koraykv Β·
Launched Gemini 3.1 Flash Live. Itβs capable of handling the nuances of live speech, like tone and interruptions, that are critical for real-world interactions. You can experience it on Gemini Live and Search Live!
Watch video
@illyism Β·
google/gemini-3.5-flash-lite: - Same cost as Gemini 2.5, 25% faster, and substantially better output - 8.6Γ faster and 94Γ cheaper than Gemini 3.6 - 6Γ faster and 22Γ cheaper than Luna, with much better output
@Marie_Haynes Β·
How did I ever live without Gemini in Chrome before? It appeared yesterday and I am finding myself using it all day long: -Help me understand this paper. -Is this a new page? There's no date on it and I can't tell. -Look at this site plus the Ahrefs / GSC tab I have open and give me some quick observations. -Here are the keywords in GSC driving traffic to this page. Find me ones that are getting impressions but the page is not optimized for. Give me specific improvements to make. -I'm on the main blog page. Make me a list of each of the blog posts linked to from here with one url per line. -How the heck do I find out which blog posts are generating revenue from organic search using GA4? For this last one I turned on Gemini Live from the Chrome sidebar and had Gemini walk me through what to do one step at a time what to do. I expect that soon it will be agentic so I can just ask Gemini to do it for me. (Claude can sometimes do this but it was struggling with GA4 today - I can relate actually.) We are moving to an era where Google is not a search engine to help me find information, but moreso an assistant to help me do things.
@JulianGoldieSEO Β·
Google just linked NotebookLM and Gemini together. And it completely fixes generic AI writing. Most AI sounds like a robot because it guesses the facts. Now you can fix that in three simple steps. First, drop your messy PDFs and notes into NotebookLM. It turns them into a private research library. Second, click the new Cinematic Video Overview button. It automatically builds a short video presentation of your notes. Third, open Google Gemini. Gemini can now read your private NotebookLM library. Ask Gemini to write five emails based only on your notes. You get expert writing with zero generic fluff. It takes ten minutes to set up. Both tools have free versions you can use right now.
Watch video
@andrewbrown Β·
I am working on my Gemini CLI Essential courses. The first video is the install video. I give it the simple task of creating a hello_world file with a print line. It took Gemini multiple minutes to complete the task, and just absolutely weird results like after it made it, deleting moving and recreating... Who the heck on the @GeminiApp is doing QA? Maybe its because its a "preview" model, but don't set it as the default if its not ready for prime time. The word start to a coding ai tool so far.
Watch video
@suraj_sharma14 Β·
Google Gemini's AI Challenges - Flash is a competent conversational model, yet it underperforms compared to Grok regarding agentic loops, even in simple instances. - Pro is a legacy model. - Veo is cost-prohibitive. SeeDance is superior. - Nanobanan Pro is outdated. GPT image generation is more effective. - Flash Lite exhibits high latency, rendering it significantly less usable. In summary, Google lacks a leading AI model for core workloads. While the chat functionality is adequate, Grok and Luna are also highly effective and demonstrate superior instruction adherence.
@JustinThomasAI Β·
I used Google Gemini's new Deep Research feature to help me with my bull debit spread options strategy. I really like that you can conduct deep research inside a notebook powered by NotebookLM. Each conversation will be saved as a source for later. Gemini used more sources than any other LLM I have used, and I also liked that it separates sources by used in this report, read but not used in this report, and then shows how it thought throughout the entire report. Maybe the best part is being able to take that report and immediately turn it into a web page, infographic, quiz, flashcards, or audio overview. The research was thorough, and what I liked the most was that it did not just agree with my strategy. It actually suggested a new strategy for me, gave me reasons why, and then told me how to implement it. Overall, very pleased.
Watch video
@rseroter Β·
A new model version, let alone a new model family/provider, is rarely a drop-in replacement. Things change from release to release. Tool handing, thinking levels, you name it. Here's a @GoogleCloudTech guide for migrating to the latest Gemini model: https://t.co/iCJ7QLSll9
@JulianGoldieSEO Β·
GOOGLE GEMINI FLASH IS QUIETLY BEATING THE βBIGβ MODEL And one hidden setting is making most peopleβs answers worse right now. What It Can Do Now: β Turn a messy whiteboard photo into a clean business growth plan β Read hidden words behind a coffee stain instead of guessing β Watch a full training video and pull exact timestamps β Rebuild charts directly inside the chat β Turn an ugly napkin sketch into a working landing page The Workflow Upgrade: β Drop in a 50-page agreement and ask it to flag penalties + auto-renewals β Upload a 2-minute voice note and get a ranked task list in seconds β Screenshot a broken automation and get the real fix without searching forums β Connect Drive, Docs, and Gmail to create docs + draft emails from one prompt The Hidden Setting: β Thinking level matters. Low = fast. High = slower, but much smarter. On high, Gemini caught two penalty clauses and a sneaky auto-renewal that low missed. Use low for quick emails. Use high when being wrong costs you money. Gemini is not just a chatbot anymore. Itβs turning messy inputs into finished work.
Watch video
@onenewbite Β·
.@antigravity @OfficialLoganK Iβm seriously considering cancelling Google AI Ultra. Itβs 4:30 PM in Malaysia and you are telling you are at capacity while the whole North America is asleep. Recently, even the Ultra subscription reduces the Opus 4.6 quota significantly and I often need rely on extra AI credits to continue working. Yea, Iβd love to use Gemini 3.1 Pro but to be honest, it sucks compared to Opus. I simply canβt trust Gemini models for serious work. After I have accepted the reality of lower quota and extra cost, I canβt even use the service reliably despite being the βUltraβ user. Really disappointed. Google/gemini is like the guy who has real potential but never grows up and delivers. And I donβt have time to wait.
@theinformation Β·
Founders need to meticulously prepare for skeptical venture capitalists when looking to raise. In the latest installment of The Gemini Edge, we look at how founders can do just that using Google Gemini to stress-test their investor pitches. By prompting the AI to act as a skeptical venture capitalist, you can prepare for Q&A, identify blind spots and uncover new risks and develop rigorous rebuttals to help you best prepare for your pitch. Supported by @GeminiApp
Watch video
@apoorv03 Β·
AI Nuggets from Google's Q2 earnings! The big takeaway is Gemini is scaling, Search is now AIβnative via AI Overviews / AI Mode / Gemini, and even generative video is scaling. ...and if you are wondering what's the difference between Gemini, AI Overviews, and AI Mode - you're not alone. I had to "AI overview" that as well lol π Google's AI by the numbers: - Geminiβ―450β―M MAUs, requestsβ―upβ―50β―%β―QoQ -- wow, this seems much larger than I'd have expected. Roughly a third of the of ChatGPT! - AI Overviews is 2 B MAUs and 5% of those click on the "AI Mode" button ~100m MAUs. - Powered by Veo3, Googleβ―Vids at ~1β―M MAUs (short video product) Some other enterprise nuggets: - tokens processed are 980β―T tokensβ―/β―month (2xβ―sinceβ―May) - 9β―M devsβ―&β―85β―k+ enterprises on Gemini (35xβ―YoY)
@glenngabe Β·
Also "inherently multilingual" -> Google launches Gemini 3.1 Flash Live, an audio model with improved tonal understanding and lower latency for real-time dialogue, watermarked with SynthID "With the 3.1 Flash Live model under the hood, Gemini Live delivers faster responses compared to the previous model and it can follow the thread of your conversation for twice as long, keeping your train of thought intact during longer brainstorms." "3.1 Flash Live is also inherently multilingual, which enables this weekβs global expansion of Search Live. With this launch, people in more than 200 countries and territories can now have real-time, multimodal conversations with Search in their preferred language." https://t.co/AOtF4RIl32
@janetmachuka_ Β·
I love how Google thought out this one! Watching a video while doing more research or asking questions on the Gemini AI button is genius. Very good at just selecting key areas in the video to watch. Love this!
@glenngabe Β·
Focus on news SEO? Heads-up, I tried it this morning and it's pretty darn good. And you can interrupt it to find out more about a story -> Gemini Live just killed boring news briefs "News summaries in Gemini Live are now deeper and more interactive. Try asking "Whatβs the latest news?" or "Catch me up on tech news" during a Live conversation to get detailed updates. You can even ask follow-up questions to dive deeper into a story." https://t.co/l2JgtwFi84
Google's AI assistant Gemini experienced one of its largest outages to date, affecting users across web, mobile, and desktop platforms worldwide. The disruption lasted for several hours, with users reporting Error 1076 and Error 1099 messages whenever they attempted to submit prompts. Google later confirmed the issue and stated that engineering teams were working to restore services. The incident sparked conversations about the growing dependence on AI tools for work, research, coding, and productivity. As AI becomes deeply integrated into daily workflows, outages like this highlight the importance of infrastructure reliability and resilience across the AI ecosystem. π¦πΌππΏπ°π²: Google Workspace Status Dashboard, TechRadar, Times of India #Google #Gemini #AI #ArtificialIntelligence #TechNews #Innovation #Orbis86
@ThePracticalDev Β·
One of the first deep dives into wiring Google Workspace MCP servers to Gemini CLI. Gmail, Drive, Calendar, People β all connected via OAuth 2.0 with working demos included. { author: xbill + @GoogleDevExpert } https://t.co/31NQGON9uH
@ThePracticalDev Β·
The Google Agent Development Kit was used to build a multimodal agent with the Gemini Live model β tested locally with the CLI, then deployed to Amazon Fargate. A step-by-step walkthrough of the environment setup, ADK verification, and the full Fargate deployment. { author: xbill + @GoogleDevExpert } https://t.co/YY6akEA1vU
@theinformation Β·
Want to understand customer sentiment better? Google Gemini can analyze thousands of customer comments in minutes, helping your team identify patterns, contradictions and root causes behind complaints. It expedites time consuming slow, error-prone manual review with a system for turning feedback into critical product and business recommendations to help decision making. Read more: https://t.co/PHwOmxB0TP Supported by @GeminiApp
Best Google Gemini tweets
Xholic studies what works in your niche, drafts posts in your voice and schedules them for the hours your audience is online.
No card needed Β· Nothing to cancel
Browse all tweet collectionsLive demo: no sign-up or login needed Β· takes about 2 minutes
Keep exploring