Open weights and open ledgers now carry the deployment decision
Gemma 4, Codex pay-as-you-go and OpenAI $122 billion round, and The Deploy Log for this week. Every price and claim read off the publisher's own page, with a call on each.
Google DeepMind
Gemma 4
DEPLOYif you run models on your own hardware
WHAT SHIPPEDGemma 4 ships in four sizes: Effective 2B, Effective 4B, 26B MoE, and 31B Dense. The 31B ranks #3 among open models on the Arena AI text leaderboard, and the 26B ranks #6. The 26B MoE activates only 3.8 billion parameters during inference.
THE CATCHThe edge models run offline on phones, Raspberry Pi, and NVIDIA Jetson Orin Nano with near-zero latency, and Android developers can prototype agentic flows in the AICore Developer Preview today.
IF YOU RUNa local code assistant or agentic workflow on a laptop GPU or workstation
SOURCEGoogle DeepMind, Gemma 4: Byte for byte, the most capable open models
OpenAI
Codex pay-as-you-go
DEPLOYif you run a small engineering team
WHAT SHIPPEDTeams on ChatGPT Business and Enterprise can add Codex-only seats with pay-as-you-go pricing and no rate limits. The annual price of ChatGPT Business drops from $25 to $20 per seat. Eligible workspaces receive $100 in credits per new Codex-only team member, up to $500 per team.
THE CATCHCodex users within ChatGPT Business and Enterprise have grown 6x since January, and more than 2 million builders now use Codex every week.
IF YOU RUNa pilot where a few engineers need Codex without a fixed seat fee
SOURCEOpenAI, Codex now offers pay-as-you-go pricing for teams
OpenAI
OpenAI $122 billion round
WATCHif you build on OpenAI APIs
WHAT SHIPPEDOpenAI closed a funding round with $122 billion in committed capital at a post money valuation of $852 billion. The company is generating $2 billion in revenue per month, and ChatGPT has more than 900 million weekly active users and over 50 million subscribers.
THE CATCHOpenAI says it is growing revenue four times faster than Alphabet and Meta did at the same stage, and enterprise now makes up more than 40% of revenue.
IF YOU RUNa product that depends on OpenAI model pricing and compute availability
SOURCEOpenAI, OpenAI raises $122 billion to accelerate the next phase of AI
The Deploy Log
Models and APIs
| DEPLOYif you extract data from documents Granite 4.0 3B Vision. Granite 4.0 3B Vision is available on Hugging Face under Apache 2.0 and leads on PubTablesV2 cropped at 92.1 and full-page at 79.3, OmniDocBench at 64.0, and TableVQA at 88.1. Hugging Face |
| DEPLOYif you post-train models in production TRL v1.0. TRL v1.0 is released with a stable core following semantic versioning and an experimental layer, and the library is downloaded 3 million times a month. Hugging Face |
| DEPLOYif you run open-vocabulary segmentation Falcon Perception. Falcon Perception is a 0.6B-parameter early-fusion Transformer that reaches 68.0 Macro-F1 on SA-Co versus 62.3 for SAM 3, and Falcon OCR is a 0.3B model scoring 80.3 on olmOCR and 88.6 on OmniDocBench. Hugging Face |
| WATCHif you work on codon optimization mRNA models 25 species. OpenMed trained 4 production mRNA language models across 25 species in 55 GPU-hours for $165, with CodonRoBERTa-large-v2 reaching perplexity 4.10 and Spearman CAI correlation 0.40. Hugging Face |
Tools and products
| DEPLOYif you pay for Vercel observability Observability Plus fee. Vercel eliminated the $10 base fee for Observability Plus for all customers and now charges only for observability events collected. Vercel |
| DEPLOYif you run delayed background jobs Vercel Queues 7-day TTL. Vercel Queues now supports a maximum message TTL and delivery delay of 7 days, up from 24 hours, while the default 24-hour TTL and 0-second delivery delay remain unchanged. Vercel |
| WATCHif you proxy to external origins CDN cache-control headers. New Vercel projects will honor cache-control headers by default when proxying requests to external origins starting April 6th, and existing projects can opt in from the project dashboard. Vercel |
| WATCHif your builds use axios axios supply chain attack. The axios npm package was compromised in a supply chain attack discovered on March 31, 2026, affecting axios@1.14.1 and axios@0.30.4, and the safe release is axios@1.14.0. Vercel |
Left out
- OpenAI acquires TBPN: the publisher's page does not show it available yet
- Gradient Labs bank agents: the publisher's page does not show it available yet
- Qwen 3.6 Plus: the publisher's page does not show it available yet
- Gemma 4 on AI Gateway: the publisher's page does not show it available yet
- GLM 5V Turbo: the publisher's page does not show it available yet
- Notion tabs: a figure in the draft could not be found on the publisher's own page
- Chat SDK Zernio adapter: the publisher's page does not show it available yet
- Nuxt MCP Toolkit: the publisher's page does not show it available yet
- Sandbox snapshot speed: a figure in the draft could not be found on the publisher's own page
- Workflow custom serialization: it did not carry a number that moved
- Marketplace resource transfers: it did not carry a number that moved
- Go backends on Vercel: it did not carry a number that moved
- gradio.Server: it did not carry a number that moved
- Spaces CLI: it did not carry a number that moved
The pattern
The deployment decision now runs on open weights and open ledgers. Gemma 4 ships in four sizes with the 31B ranking #3 among open models on the Arena AI text leaderboard, while Granite 4.0 3B Vision leads on PubTablesV2 cropped at 92.1 and TableVQA at 88.1 under Apache 2.0, and Falcon Perception reaches 68.0 Macro-F1 on SA-Co versus 62.3 for SAM 3. These open models run offline on phones, Raspberry Pi, and NVIDIA Jetson Orin Nano, and TRL v1.0 is downloaded 3 million times a month. The cost of training them collapses: OpenMed trained 4 production mRNA language models across 25 species in 55 GPU-hours for $165. OpenAI closed a funding round with $122 billion in committed capital at a post money valuation of $852 billion, generating $2 billion in revenue per month, and ChatGPT has more than 900 million weekly active users and over 50 million subscribers. Teams on ChatGPT Business and Enterprise can add Codex-only seats with pay-as-you-go pricing and no rate limits, and the annual price of ChatGPT Business drops from $25 to $20 per seat. More than 2 million builders now use Codex every week.
The infrastructure layer is removing fixed fees and extending time horizons. Vercel eliminated the $10 base fee for Observability Plus and now charges only for observability events collected. Vercel Queues now supports a maximum message TTL and delivery delay of 7 days, up from 24 hours. New Vercel projects will honor cache-control headers by default when proxying requests to external origins starting April 6th. The axios npm package was compromised in a supply chain attack discovered on March 31, 2026, affecting axios@1.14.1 and axios@0.30.4, and the safe release is axios@1.14.0. Open weights, open ledgers, and open infrastructure now carry the deployment decision together.