Frontier models ship fast while the tooling layer consolidates around them
GPT-5.2-Codex and Gemini 3 Flash, and The Deploy Log for this week. Every price and claim read off the publisher's own page, with a call on each.
OpenAI
GPT-5.2-Codex
DEPLOYif you run long-horizon coding or security work
WHAT SHIPPEDGPT-5.2-Codex is a version of GPT-5.2 optimized for agentic coding in Codex, with improvements on long-horizon work through context compaction, stronger performance on large code changes like refactors and migrations, improved Windows performance, and significantly stronger cybersecurity capabilities. It achieves state-of-the-art performance on SWE-Bench Pro and Terminal-Bench 2.0.
THE CATCHOpenAI is pairing the release with a pilot trusted access program for vetted security professionals, and the model does not reach High cyber capability under the Preparedness Framework, so the strongest defensive use cases are gated behind an invite-only program.
IF YOU RUNa security team doing vulnerability research or authorized red-teaming, you can express interest in the trusted access pilot
SOURCEOpenAI, Introducing GPT-5.2-Codex
Google DeepMind
Gemini 3 Flash
DEPLOYif you build interactive apps or agentic workflows
WHAT SHIPPEDGemini 3 Flash offers Pro-grade reasoning at Flash-level latency, efficiency, and cost. It scores 90.4% on GPQA Diamond, 33.7% on Humanity's Last Exam without tools, 81.2% on MMMU Pro, and 78% on SWE-bench Verified, outperforming Gemini 2.5 Pro and Gemini 3 Pro on that coding benchmark. It uses 30% fewer tokens on average than 2.5 Pro and is 3x faster.
THE CATCHGemini 3 Flash is priced at $0.50 per 1M input tokens and $3 per 1M output tokens, and it is now the default model in the Gemini app and AI Mode in Search, so the free tier gets the same model developers pay for.
IF YOU RUNa team already on Gemini 2.5 Pro, you can switch to Gemini 3 Flash and get better SWE-bench Verified performance at a fraction of the cost
SOURCEGoogle DeepMind, Gemini 3 Flash: frontier intelligence built for speed
The Deploy Log
Models and APIs
| DEPLOYif you generate product catalogs or brand visuals ChatGPT Images. The new ChatGPT Images model, powered by GPT Image 1.5, generates images up to 4x faster, makes precise edits while keeping details intact, and image inputs and outputs are now 20% cheaper in GPT Image 1.5 compared to GPT Image 1. OpenAI |
| WATCHif you audit or debug AI agents Gemma Scope 2. Gemma Scope 2 is an open suite of interpretability tools for all Gemma 3 model sizes from 270M to 27B parameters, produced by storing approximately 110 Petabytes of data and training over 1 trillion total parameters. Google DeepMind |
| WATCHif you assess model risk before deployment GPT-5.2-Codex System Card. GPT-5.2-Codex is very capable in the cybersecurity domain but does not reach High capability on cybersecurity, and it is being treated as High capability on biology with the corresponding suite of safeguards. OpenAI |
Tools and products
| DEPLOYif you manage large-scale migrations or expired pages Bulk redirects GA. Vercel's bulk redirects are now generally available via UI, API, or CLI without a new deployment, allowing up to one million static URL redirects per project, with Pro plans including 1,000 redirects and Enterprise including 10,000. Vercel |
| DEPLOYif you manage services from Vercel dashboard Vercel Agent for Marketplace. Vercel Agent in the Dashboard can now interact with installed Marketplace integrations from Neon, Supabase, Dash0, Stripe, Prisma, and Mux, with tools exposed by providers available automatically and authentication handled by Vercel. Vercel |
| DEPLOYif you use Cline for daily coding workflows Cline on AI Gateway. Cline now runs on the Vercel AI Gateway, and over a week of live A/B testing, P99 streaming latency improved by 10-14% across Cline's most-used models while API error rates dropped by 43.8%. Vercel |
| DEPLOYif you build multi-tenant platforms on Vercel Dynamic URL prefixes. Vercel now supports dynamic URL prefixes for multi-tenant platforms, allowing a single project to be backed by tens of thousands of domains with prefixes like tenant-123---project-name-git-branch.yourdomain.dev. Vercel |
Left out
- ChatGPT app submissions: the publisher's page does not show it available yet
- Mistral OCR 3: the publisher's page does not show it available yet
- AWS databases on Vercel: the publisher's page does not show it available yet
- Vercel Knowledge Base: the publisher's page does not show it available yet
- TLD price reductions: a figure in the draft could not be found on the publisher's own page
- OpenAI Academy for News: the publisher's page does not show it available yet
- Observability CSV/JSON export: it did not carry a number that moved
- vercel.ts: it did not carry a number that moved
- Gemini 3 Flash on AI Gateway: its own name for the thing was too long to print
The pattern
This week the frontier moved on two axes at once. OpenAI shipped GPT-5.2-Codex with state-of-the-art SWE-Bench Pro and Terminal-Bench 2.0 scores, and Google DeepMind shipped Gemini 3 Flash at $0.50 per 1M input tokens with 78% on SWE-bench Verified, beating Gemini 3 Pro on that benchmark. The deployment decision is no longer about capability alone; it is about which model fits the workflow, because both are available now and both undercut their predecessors on cost or speed.
The tooling layer is consolidating around Vercel. AWS databases, Gemini 3 Flash, Cline, bulk redirects, vercel.ts, and Marketplace agent chat all landed in the same week, which means a team on Vercel can now provision production databases, route frontier models, and manage redirects without leaving the dashboard. The pattern is that infrastructure vendors are absorbing the integration work so developers spend less time wiring and more time shipping.