Best tweets about Gemma
Find the best tweets about Google Gemma, including open model releases, benchmarks, fine-tuning, local deployment, safety, and developer experiments.
Google Gemma model releases, research, benchmarks, fine-tuning, deployment, and comparisons, excluding gemstones and unrelated names.
Best Gemma tweets
Ranked 01–50
@adrgrondin ·
Google’s Gemma 4 E2B running on-device on iPhone 17 Pro Gemma 4 is built from the same research as Gemini 3, has image understanding capabilities and can reason if needed Running at ~40tk/s with MLX optimized for Apple Silicon
Watch video
@UnslothAI ·
You can now fine-tune Gemma 4 with our free notebooks! 🔥 You just need 8GB VRAM to train Gemma 4 locally! Unsloth trains Gemma4 1.5x faster with 50% less VRAM. GitHub: https://t.co/aZWYAtakBP Guide: https://t.co/NBwKoFH2lp Gemma-4-E4B Colab: https://t.co/JjpCQgWEpL
@MaziyarPanahi ·
Gemma 4 just dropped. I had it captioning video in real-time within an hour. Running locally on a MacBook. No cloud. No API. Real-time scene understanding. Oh and SAM3 is segmenting every object in the same frame. Same laptop.
Watch video@adrgrondin ·
Google's Gemma 4 26B A4B running on MacBook Pro M5 Max 100% native Swift Achieving over 100tk/s with Apple MLX framework
Watch video@MaziyarPanahi ·
Gemma 4 analyzes the video. Generates key questions. Calls Falcon Perception. "Find all the people." 156 found. "Detect only white cars." 8 found. A 26B model is running agentic multi-QA vision orchestration. The models are running locally on a MacBook with MLX. No API.
Watch video
@0xSero ·
As promised! Gemma-4-21B-REAP is out! Results are great it held up really well and actually gained accuracy on reasoning tasks. MLX & GGUF bros do you thing! This should fit on as little as 12GB of vram with some context, or 16GB with full context https://t.co/5x4qCDuZ5b
@sundarpichai ·
DiffusionGemma is an open, experimental model that brings our text diffusion research to Gemma 4. It’s a racehorse 🏇achieving up to 4x faster inference by generating entire blocks of text simultaneously vs predicting token-by-token (word-by-word) output!
Watch video
@rasbt ·
Flagship open-weight release days are always exciting. Was just reading through the Gemma 4 reports, configs, and code, and here are my takeaways: Architecture-wise, besides multi-model support, Gemma 4 (31B) looks pretty much unchanged compared to Gemma 3 (27B). Gemma 4 maintains a relatively unique Pre- and Post-norm setup and remains relatively classic, with a 5:1 hybrid attention mechanism combining a sliding-window (local) layer and a full-attention (global) layer. The attention mechanism itself is also classic Grouped Query Attention (GQA). But let’s not be fooled by the lack of architectural changes. Looking at the benchmarks, Gemma 4 is a huge leap from Gemma 3. This is likely due to the training set and recipe. Interestingly, on the AI Arena Leaderboard, Gemma 4 (31B) ranks similarly to the much larger Qwen3.5-397B-A17B model. But as I discussed in my model evaluation article, arena scores are a bit problematic as they can be gamed and are biased towards human (style) preference. If we look at some other common benchmarks, which I plotted below, we can see that it’s indeed a very clear leap over Gemma 3 and ranks on par with Qwen3.5 27B. Note that there is also a Mixture-of-Experts (MoE) Gemma 4 variant that is slightly smaller (27B with 4 billion parameters active. The benchmarks are only slightly worse compared to Gemma 4 (31B). I omitted the MoE architecture in the figure below because the figure is already very crowded, but you can find it in my LLM Architecture Gallery. Anyways, overall, it's a nice and strong model release and a strong contender for local usage. Also, one aspect that should not be underrated is that (it seems) the model is now released with a standard Apache 2.0 open-source license, which has much friendlier usage terms than the custom Gemma 3 license.
@demishassabis ·
Excited to launch Gemma 4: the best open models in the world for their respective sizes. Available in 4 sizes that can be fine-tuned for your specific task: 31B dense for great raw performance, 26B MoE for low latency, and effective 2B & 4B for edge device use - happy building!
@Jilles ·
Gemma 4 is now on Cloudflare Workers AI. Vision, tool calling, reasoning and a 256k context window… Here’s a simple TanStack + Workers AI compliments app. 4 compliments because it’s Gemma 4.
Watch video
@RoundtableSpace ·
GEMMA 4 26B IS NOW RUNNING LOCALLY ON A MACBOOK M1 AT JUST 13GB. And somehow it still keeps the full context window while pushing 20 to 40 tokens per second.
Watch video
@ai_for_success ·
You can run Google new Gemma 4 on mobile easily. I am using Gemma 4 version E2B on my Pixel 10 Pro. Here is all you need to do: - Go to the App Store and install Google AI Edge Gallery. If you already have it, just update it. - From there, you can install the model directly and start using it. Still exploring what this small but powerful model can do :)
Watch video
@TheTuringPost ·
4 reasons Gemma 4’s architecture runs efficiently on your hardware: 1. Local + global attention structure 4 or 5 local layers + 1 final global layer to preserve the context understanding 2. Special optimizations for global attention: - 8 query heads per KV head in Grouped Query Attention (GQA) - Doubling the key dimensionality for better capacity - Keys = values to reduce memory - p-RoPE (partial RoPE) to stabilize long-context behavior 3. Image processing pipeline that is more aware of images’ shapes, resolution, compute budget and general precision. For this, Gemma 4 uses 2D RoPE + adaptive resizing with padding + different token budgets 4. The smaller E2B and E4B variants rely on per-layer embeddings that are stored separately in cheaper memory -> they become exceptionally efficient for on-device use. Explore how Gemma 4 works and why many are now switching to Gemma 4 in their OpenClaw here: https://t.co/OQTFSetgC0
@osanseviero ·
Gemma 4 is here! 🧠 31B and 26B A4B for models with impressive intelligence per parameter 🤏E2B and E4B for mobile and IoT 🤗Apache 2.0 🤖Base and IT checkpoints available Available in AI Studio, Hugging Face, Ollama, Android, and your favorite OS tools 🚀Download it today!
@FrameworkPuter ·
Google's new Gemma 4 is excellent, and the 26B MoE version is likely the best model to run on a 32GB Framework Desktop. It's fast, smart, and also great for tool calling if you use it with @openclaw or other local agent platforms.
Watch video
@fahdmirza ·
💥 Gemma 4 E2B Uncensored is here — and it answers everything ♠ abliteration removes refusal behavior without touching model intelligence 🔬 🔹Only 3 refusals out of 686 test prompts — down from 98% 🔹KL divergence of 0.346 — model quality essentially unchanged 🔹Norm-preserving biprojection keeps weight magnitudes intact 🔹Running locally via Ollama — no cloud, no API, no censorship layer 🔹Built for AI safety research, red teaming and alignment study 🔥 Watch the full video below 👇
@kimmonismus ·
Google DeepMind released new Gemma 4 QAT models that make the model family much more efficient for local, on-device use. Using Quantization-Aware Training, the models are trained with compression in mind, which reduces memory needs while preserving more quality than standard post-training quantization. The release includes support for the popular Q4_0 format and a new mobile-specialized quantization format. Gemma 4 E2B can now run with around 1GB of memory (!), and the text-only version can even require less than 1GB (!). That makes local AI on phones, laptops, edge devices, and consumer GPUs far more practical. Really cool to see.
@_philschmid ·
Gemma 4 is here! 4⃣Our most capable, agentic open model, built on the same research as Gemini 3. ✨ Reasoning. Multimodal. Four sizes (2B to 31B). Base + Instruct. Released under Apache 2.0. Runs on your phone, laptop, or servers. 🧵↓
@ArtificialAnlys ·
Google has released Gemma 4, a new family of multimodal open-weight models including Gemma 4 E2B, Gemma 4 E4B, Gemma 4 31B and Gemma 4 26B A4B @GoogleDeepMind’s new Gemma 4 family introduces four multimodal models supporting text, image, and video inputs. We evaluated Gemma 4 31B (dense) and Gemma 4 26B A4B (MoE), both with a 256k context window, while the other two smaller models support up to 128k. With 31B and 26B parameters respectively, both evaluated models can run on a single H100. On GPQA Diamond, our scientific reasoning evaluation, Gemma 4 31B (Reasoning) scores 85.7%, the second highest result we have recorded for an open-weights model with fewer than 40B parameters, just behind Qwen3.5 27B (Reasoning, 85.8%). It reaches this score using only ~1.2M output tokens, fewer than Qwen3.5 27B (~1.5M) and Qwen3.5 35B A3B (~1.6M). Gemma 4 26B A4B (Reasoning) scores 79.2%, ahead of gpt-oss-120B (high, 76.2%) but behind Qwen3.5 9B (Reasoning, 80.6%). We are now running the Artificial Analysis Intelligence Index on all four Gemma 4 models and will share a full update once those results are complete.
@natolambert ·
Google dropped 4 different Gemma open-weight models! I'm most excited that they're finally adopting a standard Apache 2.0 open source license. This'll massively boost adoption. The standard of better licenses was set by mostly Chinese open model labs, and now labs in the U.S. companies are following suit. The models are really like 31B dense, 26B-4B active MoE, 8B, 5B dense (called smaller for some reason). Base models too. Good sizes for tinkering, some local uses, and research (8/5B). 30B is particularly a great size range for building useful tools (which is why we made Olmo 3 that size too). Gemini doesn't release bad models so I'm excited to try these! Congrats Googlers.
@emollick ·
I am impressed by Gemma 4, there’s a lot of power for an on-device model at fast speeds. But I am not convinced you can get real agentic workflows out of a small model on device. So much depends on model judgement, self-correction, and accuracy. Small models are too weak there.
@andrewdfeldman ·
@cerebras is now running @GoogleDeepMind's Gemma 4 - the leading open-weight multimodal model - at 1,851 tokens per second in public preview. This is 35x faster than a typical GPU endpoint. Cerebras speed also translates into world class latency - Gemma 4 on Cerebras returns its first answer token inclusive of reasoning in 1.5 seconds, making Cerebras the only provider that lets Gemma 4 be used in real-time settings. This is the power of wafer scale.
@kimmonismus ·
To explain why I consider Gemma 4 a bigger release than most people realize. This is a big deal because models like Gemma 4 E4B can run directly on devices, bringing powerful AI (even a 2B model ~60% on MMLU Pro) to phones, laptops, and edge systems without relying on the cloud, making it faster, cheaper, and more private. At the same time, their strong performance on everyday benchmarks shows that useful intelligence no longer requires massive infrastructure. We’re entering a phase where capable AI becomes ubiquitous, embedded into everyday devices rather than locked inside data centers. And the jump from Gemma 3 to Gemma 4 further demonstrates how well such small models will perform in another 12 months. Today they have GPT-4o levels. In a year, perhaps GPT-5 levels. We will see.
@xenovacom ·
NEW: Google releases Gemma 4, their most capable open models yet! 🤯 Apache-2.0, multimodal (text, image, and audio input), and multilingual (140 languages)! They can even run 100% locally in your browser on WebGPU. Watch it describe the Artemis II launch! 🚀 Try the demo! 👇
Watch video@_philschmid ·
Gemma 4 is now available in the Gemini API and Google AI Studio. Use `gemma-4-26b-a4b-it` and `gemma-4-31b-it` with the same `google-genai` sdk as Gemini. 📝 Text generation with generate_content . 🧭 System instruction + Function Calling example. 🖼️ Image understanding example. 🔎 Google Search grounding with source citation.
@goyalshaliniuk ·
Google just dropped Gemma 4 in three distinct flavors - the efficient E2B/E4B tiny models, the powerhouse 31B Dense, and the incredible 26B MoE, each built for different use cases. The compact E2B/E4B models leverage Per-Layer Embeddings (PLE) to stay lightweight and fast, while the full 31B Dense model stacks deep FFNN layers for heavy-duty reasoning and complex tasks. The real showstopper is the 26B Mixture-of-Experts — smartly activating only 4B parameters per token, delivering flagship-level intelligence at a fraction of the computational cost.
@HarveenChadha ·
to run gemma 4 on your phone without any login or internet: - download Google AI Edge Gallery - select Gemma 4 E2B/E4B - thats it I wish they give support to add other edge models in future like locallyAI here is a video of gemma 4 E4B in action at 13 tps on iphone 15 pro max
Watch video@natolambert ·
Gemma 4 looks great on paper, just what many people need, but mostly I'm focused on the questions around what makes an open model succeed in the long term. We desperately need to build a field of research on "what makes a finetunable model" if we want the open model economy to succeed. Too much of that research happens behind closed doors of companies. Plus, 240 days after The ATOM Project was released, I share reflections on the recovery of American models, and what that means for the next phase. https://t.co/Kiex6catdp
@OpenRouter ·
Gemma 4 from @GoogleDeepMind has hit 2.5B tokens so far: Two flavors: 31B dense, & 26B MoE (running 5x faster than the dense version now!) 256K context • native function calling • multimodal • configurable thinking • 140+ languages Try the 26B MoE: https://t.co/5RIBAxvexT
@hugobowne ·
Gemma 4 came out yesterday so @canyon289 & I have updated our 10 hours of free workshops so you can build AI products with this incredible family of models today. In this repo, you'll find lesson, videos, code, notebooks, and reusable templates to build local privacy-preserving AI-powered applications, agents, MCP, image agents, eval harnesses, and more: https://t.co/gLMpPQJqzz More details in thread 👇
@RoundtableSpace ·
Google's Gemma 4 is now running fully on-device on an iPhone 17 Pro. Same research base as Gemini 3. Image understanding. Reasoning. 40 tokens per second on Apple Silicon. No internet. No cloud. A Gemini-class model in your pocket.
Watch video@ai_for_success ·
Google Just Dropped Gemma 4 this is most capable open model family from Google built for reasoning and agentic workflows. TLDR: - Available in 4 Sizes E2B , E4B for edge/on-device (phones, Raspberry Pi, Jetson Nano), 26B MoE and 31B Dense for single cloud GPU or consumer workstation - Advanced multi-step reasoning, stronger math and instruction-following - Native function-calling, JSON output, system instructions for agentic workflows - Vision and video across all models (OCR, charts, variable resolution) - Audio input on E2B and E4B for speech recognition - High-quality offline code generation - 140+ languages natively - 128K context on edge models, up to 256K on larger ones - Apache 2.0, full control, no restrictions - 26B and 31B outperform much larger models on Arena - Built on the same research stack as Gemini 3
@AIatAMD ·
1️⃣ Gemma 4️ dropped from @Google. 2️⃣ 256K context. 4️ sizes. Dense. Mixture of Experts. 3️⃣ Day 0 support on AMD Instinct, Radeon, Ryzen AI and Ryzen AI Max. 4️⃣ Deploy now on @sgl_project, @lmstudio, @vllm_project, @ollama. For all the details ↓ https://t.co/QXTom1Co4X
@aiDotEngineer ·
🆕Gemma, DeepMind's Family of Open Models https://t.co/DsOgFztNqy In the first ever public talk after the Gemma 4 launch, @osanseviero recaps how @GoogleDeepMind has packed the most capability-per-bit open source LLMs in the world (now over 500M downloads), dramatically pushing the Pareto frontier from 31B all the way down to (effectively) 2B, runnable on phones (@adrgrondin), comes with SOTA multimodal capabilities from object detection/bounding boxes to multilingual vision understanding, and FULLY FINE TUNABLE (with a -ton- of usecases from cancer therapy to LLM safety)
@burkov ·
Gemma 4 Technical Report: "We introduce Gemma 4, a new generation of open-weight, natively multimodal language models in the Gemma model family. Designed to advance compute efficiency and reasoning, the Gemma 4 model suite features dense and Mixture-of-Experts architectures, ranging from 2.3B to 31B parameters. Alongside improved vision and audio encoders for all model sizes, we propose a unified, encoder-free architecture for our 12B model, which ingests raw audio and image patches. Furthermore, we integrate a thinking mode, enabling Gemma models to generate reasoning traces prior to responding. We improve inference speed, memory, and compute efficiency, as well as long-context abilities through critical design choices. Gemma 4 establishes a leap in performance across STEM, multimodal, and long-context benchmarks, and rivals larger, frontier open models in human-rated tasks." https://t.co/Hz1Y2qPqJ3
@Marktechpost ·
Defeating the ‘Token Tax’: How Google Gemma 4, NVIDIA, and OpenClaw are Revolutionizing Local Agentic AI: From RTX Desktops to DGX Spark Google’s latest additions to the Gemma 4 family introduce a class of small, fast, and omni-capable models built explicitly for efficient local execution across a wide range of devices. Optimized in collaboration with NVIDIA, these models scale effortlessly from the Jetson Orin Nano edge AI modules to GeForce RTX PCs, workstations, and the DGX Spark personal AI supercomputer. Think of the Gemma 4 family as a high-performance engine for your local AI agents. Spanning E2B, E4B, 26B, and 31B variants, these models are designed for efficient deployment anywhere. They natively support structured tool use (function calling) for agents and offer interleaved multimodal inputs, meaning developers can mix text and images in any order within a single prompt...... Full analysis: https://t.co/EY5JLVcoxV
@0xsachi ·
I set up and ran the new Google Gemma @googlegemma model with less than ~10 minutes of work and a Macbook Air. Use unsloth @UnslothAI , instructions are here: https://t.co/Kmw93MbAxa Pick the specific gemma model, I used E2B. It browses the internet-- this is awesome.
@aiedge_ ·
Gemma 4 running on an iPhone with just ~500 MB of RAM! How to install Gemma on iPhone (100% offline, no subscription): Step 1. Download the app Search "AI Edge Gallery" in the App Store, or go to the direct listing (App ID: 6749645337). It's free, made by Google. Step 2. Pick your model size Open the app, tap the Models section, and choose E2B or E4B depending on your device. iPhone 15 Pro or newer is recommended for reliable performance, the A17 Pro's Neural Engine handles inference far more efficiently than older chips. Older iPhones can still run E2B, just expect slower responses and more battery drain. Step 3. Download over Wi-Fi The model itself is roughly 1.5-2GB. The app also needs a few minutes to optimize it for your specific chip after downloading. Don't close the app during this step. Step 4. Start chatting Once it's ready, prompt it like any other AI chat. Try Thinking Mode first, it shows the model reasoning step by step, which is genuinely one of the more impressive demos of what offline AI can do right now.
@rseroter ·
Gemma 4! Our most intelligent family of open models, with a commercially-permissive Apache 2 license, that you can use on the server, edge, or desktop. It’s a reasoning model that’s multimodal (including audio!) and supports tool use. Blog: https://t.co/BCTWpOpWwn Available on cloud: https://t.co/PPo1OByCTu Open source take: https://t.co/2Paqi1bCFz
@nathanhabib1011 ·
GEMMA 4 IS HERE > audio, text and image modalities > multilingual > Reasoning > optimized for on device > agentic capabilities > APACHE LICENSE An on device agentic coding powerhouse. Congrats to the @GoogleDeepMind team. The model is on par with qwen3.5-27B on GPQA while being multimodal. GPQA is a knowledge and reasoning benchmark, confirming the super strong reasoning capabilities of the model 🔥 DETAILED BENCHMARKS BELOW 👇
@ModelScope2022 ·
Meet Gemma 4! 🔥 Google DeepMind's latest open-weight family scales from ultra-mobile to server-grade. 📄 Apache 2.0. ⚡ E2B/E4B: PLE-optimized for mobile/browser, video + audio native, 128K ctx 🧠 26B A4B MoE: 26B total, 4B active per token, 256K ctx 📚 31B Dense: 89.2% AIME 2026, 2150 Codeforces ELO, 256K ctx 🛠️ All models: text + image, thinking mode, function calling, 140+ languages Weights are live on ModelScope. Time to ship! https://t.co/SjlbtObnCK
@shawnchauhan1 ·
Google's Gemma 4 just hit 2M downloads in days and ranked #3 among all open models globally. That's not a research milestone. That's a distribution milestone. The models developers default to become the infrastructure everything else is built on. For two years, that infrastructure has been almost entirely Chinese - Qwen, DeepSeek, Kimi, GLM. Gemma 4 is the first credible US challenger at the frontier open-weight tier. The race isn't about benchmarks. It's about who sets the defaults before the defaults harden.
@JulianGoldieSEO ·
𝗢𝗽𝗲𝗻𝗖𝗹𝗮𝘄 + 𝗚𝗲𝗺𝗺𝗮 𝟰 𝗴𝗶𝘃𝗲𝘀 𝘆𝗼𝘂 𝗮 𝗳𝘂𝗹𝗹𝘆 𝗹𝗼𝗰𝗮𝗹 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁 𝘆𝗼𝘂 𝗰𝗼𝗻𝘁𝗿𝗼𝗹 𝗳𝗿𝗼𝗺 𝗮 𝗧𝗲𝗹𝗲𝗴𝗿𝗮𝗺 𝗺𝗲𝘀𝘀𝗮𝗴𝗲 𝘄𝗶𝘁𝗵 𝘇𝗲𝗿𝗼 𝗔𝗣𝗜 𝗰𝗼𝘀𝘁𝘀. No cloud. No data leaving your machine. Nothing to pay after setup. Here's the full setup in 3 steps: → Go to https://t.co/493GbXWz04. Download and install. Make sure it's fully updated. → Open terminal. Type: ollama pull gemma4. Confirm it works with: ollama run gemma4 → Install OpenClaw. Open config. Point model endpoint to localhost:11434. Set model name to gemma4. Restart. Done. Every message you send through OpenClaw now routes through Gemma 4 on your own machine. Live demo: Sent one Telegram message. "Build me an SEO calculator in HTML with traffic and conversion inputs." OpenClaw passed it to local Gemma 4. Generated the full HTML and JavaScript. Wrote the file to my machine. Opened in browser. Zero cloud involved. Zero cost. Hardware tip: Start with the 26B model. It only activates 4B parameters during inference so it runs fast even on a consumer GPU. Use the 256,000 token context window. Paste your entire codebase in one shot. No chunking needed. Build skills as you go. Every repeatable workflow you solve, tell OpenClaw to save it as a skill. The whole setup gets more useful for your specific work over time.
Watch video
@0xTommyThomas ·
gemma 4 just casually destroyed every model on this leaderboard except opus 4.6 and gpt-5.2 31 billion parameters. $0.20 per run. runs locally on your machine. a year ago you needed an API key and a budget to access models this good. now you download one and run it yourself. the only models it couldn't beat cost 10 to 180x more to run. this is the part most people miss: the gap between "free and open" and "expensive and closed" is shrinking every quarter. if you're building anything with AI right now, the leverage available to a single person with a laptop has never been this high. the playbook where you need a big company's infrastructure to compete is dying in real time.
@alvarobartt ·
🤗 Welcome Gemma 4 to @huggingface! @GoogleDeepMind just shipped Gemma 4 in 4 sizes, 3 architectures, multimodal input (text + image + audio), built-in thinking, and much more, but everything under Apache 2.0!
@TheGeorgePu ·
Google just put local AI on your Mac. Run Gemma offline, no internet, no account. Sounds like openness winning. Read the fine print. You can only run Google's models, in Google's app, through Google's gallery. The weights are Apache 2.0. The experience is a walled garden. Open weights and open system are not the same thing. If you can't swap the model, it's not really 'open'. The default path is the whole game. And the default path still runs through them.
@JulianGoldieSEO ·
𝗢𝗽𝗲𝗻𝗖𝗹𝗮𝘄 + 𝗚𝗲𝗺𝗺𝗮 𝟰 𝗴𝗶𝘃𝗲𝘀 𝘆𝗼𝘂 𝗮 𝗳𝘂𝗹𝗹 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁 𝘆𝗼𝘂 𝗰𝗼𝗻𝘁𝗿𝗼𝗹 𝗳𝗿𝗼𝗺 𝗮 𝗧𝗲𝗹𝗲𝗴𝗿𝗮𝗺 𝗺𝗲𝘀𝘀𝗮𝗴𝗲 𝘄𝗶𝘁𝗵 𝘇𝗲𝗿𝗼 𝗔𝗣𝗜 𝗰𝗼𝘀𝘁𝘀. No cloud. No subscriptions. Nothing leaves your machine. Here's the full setup in 3 steps: → Go to https://t.co/493GbXW1aw. Download and install. Update to the latest version. → Open terminal. Type: ollama pull gemma4. Choose your size based on your hardware. → Install OpenClaw. Point the model endpoint to localhost:11434. Set model name to gemma4. Restart. Done. Every request now routes through Gemma 4 on your own machine. Live demo: Sent one Telegram message. "Build me an SEO calculator in HTML with traffic and conversion inputs." OpenClaw passed it to local Gemma 4. Full working calculator written and saved to my machine. Opened in browser. Zero cloud involved. Hardware tip: Start with the 26B mixture of experts model. It only activates 4 billion parameters during inference so it runs fast even on a consumer GPU. The 26B beats models 20 times its size on benchmarks. Apache 2.0 license. No usage caps. Full commercial freedom. Build whatever you want on top of it.
Watch video
@ThePracticalDev ·
Gemma 4 is Apache 2.0, natively multimodal, and accessible right now via the Gemini API. This dev walks through the 31B dense model vs. the 26B MoE architecture, chain-of-thought reasoning via the Thoughts toggle, and how one click in AI Studio exports production-ready TypeScript, Python, Go, or cURL. { author: @DynamicWebPaige + @GoogleAI } https://t.co/SuY0lYSFfb
@Amank1412 ·
Google’s Gemma 4 just leveled up now running about 2× faster. Tested on a 26B model powered by an NVIDIA RTX PRO 6000. The real difference shows when you compare: → regular inference → vs multi-token prediction Same model, totally different speed game.
Watch video@ThePracticalDev ·
Gemma 4 ships as 4 distinct models including a 26B MoE that activates only 4B parameters per token at inference. This dev tested all four on Cloud Run with scale-to-zero GPU, and the numbers show the 26B cold-starts faster than the 2B downloaded from HuggingFace. { author: Daniel Gwerzman + @GoogleDevExpert } https://t.co/mqFKjZqeaH
Best Gemma tweets
Xholic studies what works in your niche, drafts posts in your voice and schedules them for the hours your audience is online.
$0 today · Cancel anytime
Browse all tweet collectionsKeep exploring