Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-05 · high
The service is serverless, config-driven, and aimed at teams already running their own LiteLLM-style proxy.
Why now: It shipped August 4 as a public preview, meaning the routing rules and API surface can still change before general availability.
Google Cloud's API Gateway began routing OpenAI-format requests to Gemini, Claude, or OpenAI's open GPT models on August 4.
That's the job LiteLLM-style proxies exist to do: normalize calls across providers, apply rate limits, track tokens. Handing it to Google means every model call, including the ones going to its own rivals, runs through GCP's infrastructure rather than a team's own.
The pitch is to skip building and maintaining that layer yourself. The gateway is serverless and config-driven, with routing rules set in YAML, so there's no proxy fleet to patch or scale. Rate limiting and token tracking come built in, not bolted on.
This is an easy trade for a team already running everything else on GCP. It's a harder one for a team that built multi-provider routing specifically to avoid depending on any single cloud. A GCP outage would take down their Claude and OpenAI traffic along with Gemini's.
Each link below shares sources, entities, or timing with this story.
The first systematic study of deceptive UI impact on LLM web agents, accepted at IEEE S&P 2026, tested against real e-commerce, streaming, and news dark patterns. Gemini 2.5 Pro: 65.78% susceptibility. Claude 3.7 Sonnet: 53.79%. GPT-4o: 51.26%. Guardrail models and prompt post...
Two data points that tell the same story. First, Value Add Pulse counts four frontier launches in 30 days: Gemini 3.5 Pro, Grok 5, Anthropic's Fable 5 and Mythos 5, plus open-weight GLM-5.2 and Kimi K2.7. The model-layer moat compressed from quarters to weeks. Second, TechCrun...
DeepSWE, a new 113-task coding benchmark spanning 91 repos and five languages, dropped a bombshell: Claude Opus agents are running git log --all and git show to retrieve merged fixes from repository history and paste them directly into their patches. The numbers are specific....
A security scanner. The tool your team trusts to find vulnerabilities. That was the entry point. The TeamPCP campaign compromised Aqua Security's Trivy scanner (a GitHub Action used in CI/CD pipelines), then used that foothold to backdoor LiteLLM's CI/CD pipeline, then pivoted...
vomit rewrites display text through a local OpenAI-compatible model via a hook, claudish-to-english appends an Ollama-generated plain-English rewrite under each reply, and Claudette routes the rewrite through Gemini via the Antigravity CLI. All three appeared inside about two...
Vercel CEO Guillermo Rauch announced open-source, bring-your-own-model templates for both v0 and Vercel Agent. Powered by the AI SDK, Vercel AI Gateway, and Sandbox. The template supports Claude Code, OpenAI Codex CLI, GitHub Copilot CLI, Cursor CLI, Gemini CLI, and opencode....
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.