Tag: LLM

Resources about LLM collected from the weekly GCP Weekly newsletter — a hand-curated collection of Google Cloud Platform news, technical articles, tutorials, videos and release notes. Each entry below links back to the original GCP Weekly issue and retains its publish date and title, giving you a focused, searchable knowledge base for this topic.

GKE Kubernetes LLM Nov. 16, 2026

Port-forwarding to a model on a private GKE GPU cluster: DNS endpoint or Connect gateway - This technical guide explores how to securely access an LLM running on a private Google Kubernetes Engine (GKE) GPU cluster using `kubectl port-forward`.

Gemma GPU LLM Nov. 16, 2026

Scale Gemma 3 RL on GKE with NeMo RL, Custom ComputeClasses, and 8-NIC RDMA - This guide demonstrates how to apply Group Relative Policy Optimization (GRPO) reinforcement learning to Google’s Gemma 3 1B model using NVIDIA NeMo RL and KubeRay on Google Kubernetes Engine (GKE).

Cloud Run GPU LLM Nov. 9, 2026

CPU vs GPU for LLM Inference on Cloud Run: A Measured Comparison with Gemma and an NVIDIA L4 - This article evaluates running open-source LLM inference on Google Cloud Run by comparing the performance of a CPU-only setup against an NVIDIA L4 GPU configuration using the Gemma 2 2B model. It reveals that while the GPU instance costs twice as much per hour, it delivers text generation over 13 times faster and is roughly six times cheaper per generated answer due to its higher throughput.

LLM TPU Nov. 9, 2026

Part 2: Beyond the GPU Default - Choosing the Right TPU Simplified - This article provides a practical guide on how to choose and purchase the right Google Cloud TPU for specific AI workloads, moving beyond the default GPU choice. It breaks down the entire selection process by analyzing current TPU lineups, regional availability, pricing models, and memory requirements for large language models.

Antigravity LLM Nov. 9, 2026

Building a LiteLLM Enterprise Gateway for Antigravity - This article explores how enterprises can build a custom AI gateway using LiteLLM on Google Cloud Run to govern and customize their use of the Antigravity coding harness. It walks through setting up the necessary middleware to handle routing, streaming, and attribution requirements that go beyond out-of-the-box governance tools.

Apigee Generative AI LLM Nov. 2, 2026

A semantic cache is easy to demo and hard to trust - This article explores the implementation and production challenges of using semantic caches with Apigee as an AI gateway for Large Language Models.

Apigee Generative AI LLM Nov. 2, 2026

Never let a request decide who pays - Token quotas, API products and cost reports in an Apigee AI gateway, and the one rule that keeps them honest: only the credential decides.

BigQuery FinOps LLM Oct. 26, 2026

How to manage costs when prompting open source LLMs from BigQuery - A practical guide to keeping your cloud budget under control when running open LLMs in BigQuery.

LLM Official Blog Partners Oct. 26, 2026

Expanding Claude availability and deployment choice on Google Cloud - Today, we’re announcing the general availability of an Anthropic-managed deployment option for Claude Models on Google Cloud. This gives developers more choice in how they run their models, helping them consolidate AI workloads on Google Cloud by pairing direct access to Anthropic’s platform alongside our existing, in-boundary Google-managed deployment option.

LLM Official Blog Oct. 26, 2026

Google Cloud expands model choice and flexibility, bringing OpenAI models to customers - Today, we are launching OpenAI models in Google Cloud’s model garden. Google Cloud is now the only hyperscale cloud provider offering access to Gemini, Claude, and GPT, alongside more than 200 first-party, third-party, and open models on a unified enterprise AI platform.

Antigravity LLM Official Blog Oct. 26, 2026

Claude Opus 5.5 and Claude Sonnet 5.5 are now available in Google Antigravity - In keeping with Google Cloud's commitment to an open ecosystem and to give developers the right model for every job, Claude Opus 5.5 and Claude Sonnet 5.5 are now available in Google Antigravity.

Apigee Generative AI LLM Oct. 19, 2026

Stop building one proxy per model - This article explores how organizations can avoid the complexity of managing separate proxies for every AI model by consolidating them into a single, OpenAI-compatible AI gateway using Apigee.

AI Hypercomputer LLM Official Blog Oct. 19, 2026

Day 0 support for Reflection’s Beam on Google Cloud: 5x faster perf with llm-d and NVIDIA Dynamo - Google Cloud has optimized running Reflection AI's Beam model for agentic workloads, using llm-d and NVIDIA Dynamo to achieve up to 5x higher throughput and 8x lower latency on A4X Max instances. These improvements come from smarter request routing, KV-cache management, and optimized GPU usage, with ready-to-use deployment configurations available for GKE.

Gemini LLM Paywall Oct. 12, 2026

Gemini 2.5 on Vertex AI Retires October 20: What to Switch To (and What It Costs) - Google Cloud is set to retire its Gemini 2.5 models on Vertex AI by October 20, 2026, requiring developers to migrate to newer 3.x alternatives. This comprehensive guide outlines the recommended model replacements, pricing changes, and potential cost impacts for different workloads.

Apigee LLM Oct. 5, 2026

Per-User LLM Spend Capping on Apigee X - This article explores a comprehensive technical solution for enforcing per-user financial spending caps on Large Language Models using Apigee X on Google Cloud.

Gemma LLM TPU Oct. 5, 2026

Gemma 4 QAT on One TPU v5e: What Runs and What Doesn't - This article explores a step-by-step guide for repacking Google's Quantization-Aware-Trained (QAT) Gemma 4 weights to efficiently serve models up to the 12B and 26B sizes on a single Google Cloud TPU v5e chip.

Gemma LLM Official Blog Startups Oct. 5, 2026

Why your startup needs open models alongside frontier APIs - With more than one billion downloads across the developer community, Gemma 4 is Google’s most capable open model family to date, with myriad applications available to diverse startup use cases.

Cloud Run LLM Sept. 28, 2026

How to build a Jev-style classifier with DiffusionGemma and vLLM - Fast, calibrated routing on Cloud Run GPUs without fine-tuning.

Generative AI LLM Machine Learning Sept. 28, 2026

How to Run Open Models in the Cloud Without Going Broke - This article explores practical strategies for deploying open-weights machine learning models on Google Cloud’s Model Garden without incurring massive expenses. It highlights cost-saving techniques such as enabling scale-to-zero, leveraging spot virtual machines, and properly right-sizing hardware limits to make advanced cloud AI affordable for developers and smaller teams.

AI JAX LLM TPU Sept. 28, 2026

Reproducing OLMo 3 7B Pre-training in MaxText: case study of large scale training on TPUs - The MaxText team successfully reproduced AI2’s OLMo 3 7B language model from scratch on Google Cloud TPUs using JAX/XLA, precisely matching the original PyTorch-on-GPU reference across pre-training and mid-training stages on all held-out evaluations. The implementation achieved up to 57.4% Model Flops Utilization (MFU) and demonstrated robust infrastructure portability by surviving mid-run cluster resizes and cross-generation TPU shifts without requiring recipe alterations.

Gemma LLM TPU Sept. 28, 2026

Google's QAT Gemma 4 26B-A4B on One TPU v6e: 15.6x the KV Cache and 1.9x the Throughput of FP8 - This article provides a step-by-step guide to efficiently serving Google's quantization-aware-trained (QAT) Gemma 4 26B-A4B model on a single Google Cloud TPU v6e chip using vLLM.

Agents LLM Sept. 14, 2026

How Google Cloud Plugins are the Ultimate Cloud Toolkit for Claude Code - Discover how Google Cloud Plugins streamline your development workflow by seamlessly integrating official Agent Skills and Model Context Protocol servers directly into Claude Code. The article provides a step-by-step guide on how to easily discover, browse, and install these powerful foundational and domain-focused bundles straight from your terminal or workspace.

Gemini LLM Sept. 14, 2026

Inside Google’s Gemini 3.8 Flash and FlashAttention-3: The Hardware Physics of 2M Tokens - This article provides an in-depth engineering analysis of how Google Cloud achieves extreme long-context capabilities (up to 2 million tokens) in its Gemini frontier models. It breaks down the mathematical, physical, and architectural mechanics—such as FlashAttention, Grouped-Query Attention, and Ring Attention—that overcome the traditional quadratic memory bottleneck of Transformers.

AI LLM TPU Sept. 14, 2026

Autonomous LLM post-training with Tunix on TPUs - This article introduces autofinetune, a project that automates the LLM post-training process using AI agents to run iterative experiments overnight on Google Cloud TPUs. It demonstrates how autonomous research loops can handle both supervised fine-tuning and reinforcement learning using tools like Tunix and Gemma, significantly reducing manual hyperparameter tuning.

AI Kubernetes LLM Aug. 31, 2026

Deploying GKE Inference Gateway Powered by llm-d: The Ultimate Guide to Model-Aware Routing on Kubernetes - How to eliminate GPU bottlenecks, unlock prefix cache hits, and achieve predictable Time-to-First-Token (TTFT) at scale.

Gemini Enterprise Agent Platform LLM Aug. 31, 2026

Grok 4.6 available now on Gemini Enterprise Agent Platform - SpaceXAI’s advanced Grok 4.6 model is now available in Public Preview as a fully managed, serverless endpoint within Model Garden on the Gemini Enterprise Agent Platform. The integration provides developers with OpenAI-compatible APIs, enterprise-grade security, and native billing while supporting complex, long-running agentic tasks.

Cloud Run Gemma LLM Aug. 24, 2026

Fine-Tuning and Deploying Gemma 4 at Scale on Gemini Enterprise Agent Platform (GEAP) and Cloud Run - A Healthcare Guide to Supervised Fine-Tuning, Knowledge Distillation, and Sub-Second GPU Inference.

Gemini Gemini Enterprise Agent Platform LLM Aug. 24, 2026

Gemini Play Smart — a 429-survival capacity ladder - This article introduces the "capacity ladder" strategy as a cost-effective alternative to purchasing Provisioned Throughput for handling `RESOURCE_EXHAUSTED` (429) errors when working with the Gemini API on Google Cloud.

BigQuery Generative AI LLM Aug. 17, 2026

Future-Proofing Your Vector Search: Swapping Open Source Models in BigQuery with Zero Friction - This article explores how to future-proof vector search workflows by integrating Hugging Face's open-source embedding models directly into Google Cloud's BigQuery using SQL.

Gemma LLM Aug. 17, 2026

How to effectively serve MTP-based Gemma 4 models for inference performance

Showing 1–30 of 302 resources

 

Latest Issues




Contact

Zdenko Hrček
Třebanická 183
Prague, Czech Republic
Phone: +420 777 283 075
Email: [email protected]