DigitalOcean Release Notes

DigitalOcean Kubernetes (DOKS)

DigitalOcean Kubernetes (DOKS) now runs a node readiness controller in public preview. The controller watches node conditions and holds NoSchedule taints on nodes until required components, including GPU drivers on GPU nodes, report healthy, which prevents pods from scheduling onto nodes that are not yet ready. DOKS deploys and manages the controller automatically.

For GPU node pools, you can also customize which GPU health metrics gate scheduling, without redeploying any components.

Choosing the right model or inference router for production means more than reading a leaderboard

It means validating any model or routing configuration on your own data using your prompts and your evaluation criteria before it ever reaches production, and comparing quality, latency, and cost in one place.

Evaluations, now available on the DigitalOcean Inference Engine, lets teams validate any model or inference router configuration on their own data before production. Run structured LLM-as-a-Judge evaluations across catalog models, fine-tuned models, BYOM imports, and router setups without stitching together a separate evaluation stack.

DigitalOcean Evaluations Capabilities

Evaluations provide everything teams need to validate model and router performance before production. LLM-as-a-Judge scoring runs across any candidate in your inference stack and returns per-item scores with judge rationale, plus latency, token, and cost tracking per run. Six pre-built metrics cover the most common evaluation needs out of the box. For teams that need full control: custom rubrics, reusable presets, MCP support, and full dataset management — all in the same platform as the inference endpoints you use in production.

Pre-Built and Custom Rubrics: Score Against Criteria That Match Your Domain

The six pre-built metrics, correctness, completeness, faithfulness, PII, toxicity, and bias, cover common evaluation needs. For specialized domains, custom rubrics let teams define their own judge instructions and scoring criteria directly in the judge prompt.

The judge evaluates responses against these criteria and returns per-item scores with rationale. Custom rubrics can also adapt the built-in correctness metric to different data formats instead of relying on a default interpretation.

Evaluation Presets: Save Configurations and Re-Run Without Rebuilding

Without saved configurations, every re-run becomes a rebuild with different judge models, parameters, or prompts, making results hard to compare.

Evaluation presets store the full configuration of a run including judge model, metrics, system prompt, and parameters, so teams can reuse them across model and router versions and compare results directly across v1, v2, and v3 fine-tunes.

MCP Support: Trigger Evaluations Programmatically

For agentic workflows and CI pipelines, evaluations cannot be a manual step in these workflows. MCP support enables evaluation jobs to be triggered programmatically from model registration events, deployment triggers, or schedules.

API and SDK endpoints are also available for teams integrating evaluations into deployment workflows.

Dataset Management: Manage Evaluation Data as a First-Class Resource

Datasets can be uploaded, versioned, reused, and deleted in a single place. Each upload creates a versioned dataset linked to evaluation runs for traceability back to the source data.

Datasets support CSV and JSONL formats up to 1GB or 1,000 rows via Console or cURL. Optional ground truth columns can be included to enable faithfulness scoring.

How to Access Evaluations

Skip the standalone eval tools. Evaluations is natively integrated into your DigitalOcean stack, so you evaluate against the same endpoints you serve on, on infrastructure we run end to end.

Evaluations supports validating any model or inference router in your stack including models from the DigitalOcean Model Catalog, Dedicated Inference endpoints, BYOM imports from Hugging Face or Spaces, and router configurations. All evaluations run against production-grade endpoints.

Evaluations supports a range of judge models, including DeepSeek-R1-Distill-Llama-70B and Qwen3-32B. Access to premium commercial models (OpenAI and Anthropic) as candidates or judges requires a tier 2 account. You can complete a pre-payment in the Console to move to tier 2 and unlock premium model access.

Billing is based on inference tokens consumed by the candidate and the judge model. Dataset and result storage is provided at no additional charge for the first 12 months.

Your inputs, outputs, and ground truth are sent to the judge model provider for scoring only. They are not stored outside DigitalOcean and not used to train models.

Full documentation, including dataset formatting requirements, preset configuration, and MCP trigger setup, is available here.

Start Evaluating Before You Ship

Model and router decisions don’t stop after launch. Evaluations give you a repeatable way to validate on your workloads, against your criteria, on the same endpoints your users hit, as your stack evolves. Run your first evaluation in the DigitalOcean Cloud Console today.

With Prompt Caching, DigitalOcean automatically recognizes repeated token prefixes and currently offers an 80% discount on cached input tokens compared to standard input pricing across supported models. See the pricing page for current rates, applicable terms, and any pricing updates. The result is lower costs and faster time-to-first-token, with no changes to your application code. Caching applies automatically, and every API response includes a cached_tokens count so you can verify your savings in real time. It is available today across a broad set of leading open models, with more models coming soon.

Learn more ->

• DigitalOcean Managed Weaviate is now in public preview and enabled for all users. Weaviate is a fully managed vector database for retrieval-augmented generation and semantic search workloads.

  Create clusters from the Vector Databases page, the /v2/vector-databases API, or doctl vector-databases. Select Weaviate as the engine and review the public preview disclaimer and legal terms in the create flow. For setup, limits, and pricing, see Managed Weaviate.

• Claude Fable 5 is available again on DigitalOcean Inference for serverless inference, Agent Development Kit, and agents. For more information, see the Available Models page.

• Prompt caching for open-source models in serverless inference chat completions and responses API is now in public preview. Open-source models cache context automatically, so you do not need to set the cache_control or prompt_cache_retention parameters.

  Prompt caching is available for the following open-source models:

  • DeepSeek V3.2
  • DeepSeek V4 Pro
  • DeepSeek V4 Flash
  • Kimi K2.5
  • Kimi K2.6
  • GLM 5
  • GLM-5.1
  • GLM-5.2
  • gpt-oss-120b
  • MiMo V2.5
  • MiMo V2.5 Pro
  • MiniMax M2.5
  • Qwen 3.5
  • Qwen3 Coder Flash

  For more information, see Use Prompt Caching.

• The Ubuntu 26.04 LTS (ubuntu-26-04-x64) base image is now available in the control panel and via the API.

• Invoice and Billing Insights CSV files now include a tag_name column that lists the tags applied to each resource. Tag information is included only for resource usage on or after 1 July 2026.

Private Droplets

Private Droplets are now generally available in all regions. Private Droplets have no public network interface and no public IP address, using VPC-only networking with automatic integration with VPC NAT gateway, VPC peering, and VPC private DNS. See the Private Droplets documentation for setup instructions and limitations.

Agent evaluation metrics

The following agent evaluation metrics are deprecated and should no longer be used:

• Tone
• Retrieved Chunk Usage
• Prompt Perplexity

Use the currently supported metrics listed in Agent Evaluation Metrics instead. To monitor deployed agent behavior outside of evaluations, use Agent Metrics and Runtime Logs.

Anthropic model availability

The following Anthropic model is now available on DigitalOcean Inference for serverless inference, Agent Development Kit, and agents:

• Claude Sonnet 5

For more information, see the Available Models page.

Agent Development Kit evaluation support

Agent evaluations support for the Agent Development Kit (ADK), previously in preview, is now removed.

To evaluate agents, use agent evaluations via the DigitalOcean Control Panel for supported agent types. To monitor ADK agent behavior, use Agent Metrics and Runtime Logs.

Insights, agent tracing, and conversation logs

Insights, agent tracing, and conversation logs are deprecated for all agents, including agents created through the Control Panel, CLI, API, and Agent Development Kit (ADK).

To monitor deployed agent behavior, use Agent Metrics and Runtime Logs instead.

Evaluations MCP server tool

The Agent Evaluations MCP server tool has been renamed to Evaluations.

Custom metrics

Custom metrics are now available for DigitalOcean Evaluations. You can define your own metrics to evaluate model behavior against criteria specific to your use case.

DigitalOcean Evaluations

DigitalOcean Evaluations is now generally available. Use Evaluations to create test cases, run evaluation datasets, and measure model performance against selected metrics.

Presets

Presets are now available for DigitalOcean Evaluations. You can save and reuse evaluation configurations, including the candidate model, system prompt, hyperparameters, judge model, and metrics.

Model Evaluations

Model Evaluations is now renamed to DigitalOcean Evaluations.

• Serverless Inference now requires a positive prepaid account balance before you can send inference requests. Usage charges are deducted from this balance, and access is suspended if it reaches $0. You can add a prepayment manually or enable auto-reload to replenish your balance automatically. For more information, see Manage Serverless Inference Prepayment.
• The DigitalOcean Control Panel now supports light and dark themes. From the profile menu in the top right corner of the control panel, you can set your theme to a light or dark appearance, or match your operating system’s appearance setting.

GLM-5.1 and GLM-5.2, the latest open-source models from Z.ai for autonomous software engineering and repository-scale reasoning, are now available through DigitalOcean Inference Engine. GLM-5.1 is optimized for long-running agentic coding workflows and can sustain complex autonomous tasks for up to 8 hours, while GLM-5.2 introduces a 1M-token context window and dual reasoning modes for large-scale code analysis and refactoring. Both models support tool calling, structured output, and multilingual applications.

Access the models now ->

You can now attach Network File Storage shares to multiple VPC networks, and expose specific subdirectories to single VPC networks by creating access points.

A share can connect to up to 10 VPC networks in the same region. Access points restrict clients to a specific path within the share, and are isolated from one another so that clients on one cannot see the directories governed by another. This lets you host multiple tenants on a single share, each scoped to their own directory. For details, see How to Create and Delete Network File Storage Access Points.

What is DigitalOcean for Codex?

As your agents are working on more complex, long-running work, they need a clean, persistent environment to keep running. Setting up a persistent remote machine by hand means creating a cloud server, configuring SSH keys, installing dependencies, and wiring everything back to your workflow. It’s a lot of infrastructure work before you write a single line of code.

Today, we’re making that easier. The DigitalOcean plugin for Codex is now available in Public Preview, letting developers create and connect Codex-ready cloud development machines in their own DigitalOcean account directly from within Codex — using natural language, with no manual setup. This means that not only can your work continue running when you step away, but with Codex in the ChatGPT mobile app you can stay in control — starting, steering, or monitoring work from wherever you are.

The DigitalOcean plugin connects your DigitalOcean account to the Codex app, letting you provision a persistent remote development environment on demand. Instead of manually spinning up a server and configuring it, you can ask Codex to do it.

The result: a DigitalOcean Droplet® that’s pre-configured with the Codex CLI, common programming language tooling (based on the codex-universal Docker image), and SSH access — so your work can keep running and stay within reach, even when you’re not at your desk.

How it works

Starting from Codex

Install the DigitalOcean plugin from the Codex plugin directory. During installation, you’ll connect your DigitalOcean account via OAuth — no API tokens to create or paste. Then prompt:

@DigitalOcean create a new remote machine

Codex will:

• Provision a new Droplet from the Codex Droplet template
• Generate and configure an SSH key on your device
• Wait for the machine to finish provisioning (and check back automatically when it’s ready)
• Return a deeplink to the Codex SSH connections page to finalize the connection

Once connected, you’re running Codex on a persistent machine in your own DigitalOcean account. You can ask Codex to set up projects, install dependencies, configure your environment, or spin up and shut down additional machines — all from conversation.

Starting from the DigitalOcean Marketplace

Already a DigitalOcean user? Find the Codex Droplet template in the DigitalOcean Marketplace, create a Droplet, then follow the link to install the plugin in Codex. After OAuth, prompt:

@DigitalOcean connect

Codex connects the Droplet to your Codex environment using the same SSH key and deeplink flow.

You can find the droplet id in the URL at cloud.digitalocean.com/droplets/

What you can do once you’re connected

After setup, your Codex session is running on a cloud machine you control. A few things you can do from there:

• Ask Codex to configure the environment, install dependencies, or set up a new project
• Use Codex’s handoff functionality to move an existing local thread to the Droplet and continue work in the cloud
• Spin up additional machines or shut down ones you no longer need
• Connect your Droplet to Codex in the ChatGPT app to start, steer, and continue work from your phone

Getting started

The DigitalOcean plugin for Codex is available today in Public Preview. Install the plugin from the Codex plugin directory or find the Codex Droplet template in the DigitalOcean Marketplace to get started.

Get started →

https://marketplace.digitalocean.com/apps/codex-universal

Single sign-on (SSO) with OIDC for DigitalOcean Kubernetes (DOKS) clusters is now in general availability. You can authenticate users to your Kubernetes clusters through an identity provider like Auth0, authentik, JumpCloud, Keycloak, or Okta, instead of using token-based authentication.

SSO is configured per cluster with an issuer URL and client ID from your identity provider. You can enable it using doctl, the DigitalOcean API, or Terraform.

The following Z.ai model is now available on DigitalOcean Inference for serverless inference, dedicated inference, Agent Development Kit, and agents:

• GLM-5.1

For more information, see the Available Models page.

A Fedora 44 (fedora-44-x64) Droplet base image is now available in the Control Panel and through the API.

Bring Real-Time Information Into Your AI Applications

AI applications and agents are only as capable as the tools, data, and systems they can access. With Server-Side Tools, now in Public Preview for DigitalOcean Inference Engine, a model can call out to search the web, read your data, call your systems, and take action all from inside a single inference request. You can enable the new tools with your existing DigitalOcean Model Access Key. No separate tool infrastructure to assemble, no new credentials, no orchestration layer to operate.

Server-Side Tools bring web search, web fetch, DigitalOcean Knowledge Bases, MCP servers, and supported Anthropic and OpenAI tools into your inference requests, each covered below.

When applications need current information such as news, documentation, or live data, models can access the web directly during inference.

Web Search: Get live answers from the web

Web Search enables retrieval of up-to-date information from the web. This enables research workflows, support experiences, and agentic applications that need to reason over recent events, changing information, or content that is not available in a model’s training data.

Web Fetch: Pull in content from URLs and documents

Web Fetch pulls in content from specific URLs or PDFs during inference. It is useful for summarizing pages, extracting structured data from documents, or pulling in reference material on demand.

Both Web Search and Web Fetch are powered by Exa. Pricing is usage-based; see the pricing page for current rates.

Web Mode: Enable web access through the model URL

Some agent frameworks only allow you to configure a model name and do not expose tool configuration. For these cases, DigitalOcean supports Web Mode, which automatically enables Web Search and Web Fetch through the model field. This gives the model access to Web Search and Web Fetch without explicitly defining tools, making it easier to integrate with agent frameworks that only allow model-level configuration.

Ground Responses in Your Own Data

For applications that need to work with their own data and systems, Server-Side Tools provides two options.

• DigitalOcean Knowledge Bases: Give your models access to retrieve relevant content from your indexed data automatically without you building a separate retrieval pipeline. Attach your knowledge base, send the request, and the model grounds its response in your content.

• MCP Servers: Connect models to your systems and services through the Model Context Protocol. MCP servers expose internal APIs, databases, and tools, allowing models to retrieve information and take actions like writing data, updating systems, or triggering workflows directly within inference requests.

Support for Anthropic and OpenAI Tools

If you are already using Anthropic or OpenAI tool conventions, those same tool definitions work within DigitalOcean Inference Engine. There is no need to rewrite your application logic or adapt to a new interface.

• Anthropic tools include: Web fetch, Tool search, Bash, Text editor, Computer use

• OpenAI tools include: Function calling, Computer use, Tool search, Apply Patch, Local shell

All tools incur token costs based on use. For the full list of supported tools, see the documentation.

Use Your Existing Agent Tooling Without Changes

Server-Side Tools also power coding agents and developer workflows. Coding assistants such as Claude Code, Codex, and other agent frameworks rely on capabilities like web search, web fetch, bash, text editing, and computer use to gather context and complete tasks. By supporting these tools directly within inference requests, DigitalOcean Inference Engine makes it easier to run coding agents and agent frameworks without managing additional tool infrastructure.

How to Access Server-Side Tools

Server-Side Tools are available today in Public Preview through your existing DigitalOcean Model Access Key. No new credentials or account changes are required, and we plan to add more tools.

To get started, specify tools as part of your inference request, or enable Web Mode through the model URL. Server-Side Tools are available through Serverless Inference, Inference Router, and Dedicated Inference. Full documentation, including request examples and supported tool configurations, is available here.