Snorkel AI has been doing fascinating work evaluating frontier models from Anthropic, Google, etc. with multi-agent evaluation systems.

As the scale and complexity of these eval workloads increased, they wanted to move beyond fragmented logs toward a single, reliable source of truth for debugging. See how they're solving these challenges →

You can now apply rate limits and usage budgets across your entire organization, and update them instantly at a team or group level instead of handling them one at a time. You can define policy conditions based on API keys, metadata, or workspaces.

This makes it easier to:

✔ Adjust budgets org-wide in one move

✔ Roll out policy changes across teams at once

✔ Create grouped or multi-condition enforcement rules

✔ Scale access safely without config overload

See how you can implement

You can now start using Opus 4.5 in production with:

✅Unified access through Anthropic

✅Guardrails for safe, compliant usage

✅Full observability across logs, latency, and spend

✅Team-level budgets and rate-limit controls

Gemini 3.0, Google’s most advanced model yet, brings major improvements in multimodal understanding, agentic behavior, and vibe-coding accuracy — making it a strong fit for AI apps, copilots, and reasoning-heavy automation.

With Portkey, teams can start using Gemini 3.0 immediately with:

✅ Unified access via Vertex AI

✅ Guardrails for safe, compliant usage

✅ Full observability across logs, latency, and spend

✅ Team-level budgets and rate-limit controls

Try it today!

GPT-5.1, the newest model from OpenAI, is now available on Portkey. GPT-5.1 brings a step-change in speed, reasoning and flexibility.

With Portkey, you can bring GPT-5.1 into production with:

✅ Unified access across providers (OpenAI + others)

✅ Enterprise guardrails for safe, compliant usage

✅ Full observability: logs, latency, spend

✅ Budget & rate-limit controls for teams & apps

Vertex AI

• Added support for Computer Use, enabling browser automation and control capabilities.

• Added support for anthropic-beta

OpenAI

• Added support for conversation and modalities parameters for enhanced context and multimodal control.

Anthropic

• Added role to assistant in chat completion stream responses for improved stream parsing and clarity.

Portkey now has a Terraform provider for managing workspaces, users, and organization resources through the Portkey Admin API. This enables you to manage:

• Workspaces: Create and update workspaces for teams and projects
• Members: Assign users to workspaces with defined roles
• Access: Send user invites with organization and workspace access
• Users: Query and manage existing users in your organization

This makes it easier to provision and manage Portkey resources as part of your existing infrastructure workflows.

Portkey is named a Cool Vendor in the 2025 Gartner® Cool Vendors™ in LLM Observability Report! True to our recognition, we are making our observability even better!

• Tool calls are now auto-detected and displayed separately in logs — making it easy to trace each call in agent workflows.

• Web search results now appear in a cleaner, structured format for faster review and debugging.

Take a look!

Add Prefix: Automatically add a configurable prefix to user inputs before sending them to the model. This is useful for enforcing context, tone, or compliance instructions across all model interactions.

Allowed Request Types: Control which request types (endpoints) are processed by your Gateway. You can set up either an allowlist or blocklist to manage the types of LLM requests permitted

You can now configure Javelin AI Security guardrails directly in the Portkey AI Gateway and check every model interaction for:

• Trust & Safety – Detect and filter harmful or unsafe content

• Prompt Injection Detection – Identify attempts to manipulate model behavior

• Language Detection – Verify and filter language in user or model responses

See how you can add Javelin to your model interactions here: https://portkey.ai/docs/integrations/guardrails/javelin