What's New in TokenTra
Stay up to date with the latest improvements, new features, and integrations. We ship fast and often.
GPT-5 & Claude Opus 4.5 Support
Full support for the latest flagship models from OpenAI and Anthropic, including GPT-5, GPT-5.1, GPT-5 Pro, and Claude Opus 4.5.
- •Complete cost tracking for GPT-5 family (GPT-5, GPT-5.1, GPT-5 Pro)
- •Claude Opus 4.5 and Claude Sonnet 4 integration
- •Updated pricing models for new tiers
- •Enhanced token counting for extended context windows
- •Automatic model version detection
Results: Stay ahead with day-one support for the latest AI models.
Gemini 3 Integration
Full support for Google's Gemini 3 family including Pro, Flash, and Ultra variants.
- •Gemini 3 Pro, Flash, and Ultra model tracking
- •Updated Vertex AI cost calculations
- •Multimodal usage tracking for Gemini 3
- •Enhanced BigQuery export integration
Advanced Cost Forecasting v2
Our most accurate forecasting engine yet, powered by transformer-based time series models.
- •95% accuracy on 30-day forecasts
- •Automatic seasonality and trend detection
- •Per-team and per-project forecasts
- •Confidence intervals with Monte Carlo simulation
- •Slack and email forecast digests
Results: Customers report 40% fewer budget surprises with our new forecasting.
o3 Reasoning Model Support
Full support for OpenAI's o3 and o3-mini advanced reasoning models.
- •o3 and o3-mini cost tracking
- •Reasoning token accounting
- •Extended thinking time cost attribution
- •Comparison analytics vs standard models
Enterprise SSO & SCIM
Enterprise-grade identity management for large organizations.
- •SAML 2.0 and OIDC single sign-on
- •SCIM 2.0 user provisioning
- •Okta, Azure AD, and Google Workspace integration
- •Just-in-time user provisioning
- •Group-based role mapping
Smart Model Routing GA
Smart Model Routing is now generally available after successful beta with 500+ teams.
- •Automatic complexity detection for incoming requests
- •Configurable routing rules based on cost/quality tradeoffs
- •A/B testing framework to validate quality isn't degraded
- •Fallback routing when primary providers are down
- •Real-time routing analytics dashboard
Results: Teams using Smart Routing see 30-50% savings on routine queries.
AWS Bedrock Integration v2
Enhanced AWS Bedrock support with Llama 3 and Mistral Large 2.
- •Llama 3.2 and Llama 3.1 model support
- •Mistral Large 2 integration
- •Amazon Titan v2 tracking
- •Cross-region cost aggregation
- •Reserved capacity cost tracking
Semantic Caching GA
Semantic caching is now generally available with 99.9% uptime SLA.
- •Semantic similarity matching (not just exact match)
- •Configurable TTL per query type
- •Cache hit rate analytics
- •Zero latency for cached responses
- •Automatic invalidation options
Results: Customers see 15-25% cost reduction on FAQ-style and repetitive queries.
Python SDK v2.0
Major update to the Python SDK with async-first architecture.
- •Native async/await throughout
- •GPT-5 and Claude Opus 4.5 support
- •LangChain and LlamaIndex integrations
- •Automatic retry with exponential backoff
- •Structured logging with OpenTelemetry
Budget Forecasting
Know when you'll exceed your budget before it happens.
- •Holt-Winters forecasting with seasonality awareness
- •"You'll exceed budget by Tuesday" predictive alerts
- •Confidence intervals on projections
- •Weekly and monthly forecast reports
Anomaly Detection Engine
ML-powered anomaly detection catches unusual spending patterns automatically.
- •Spending spikes (Z-score > 3)
- •Unusual model usage patterns
- •Off-hours activity detection
- •New cost centers appearing
- •Rapid growth detection
Results: We analyze your historical patterns and alert you when something looks unusual—before it becomes a $10K surprise.
Google Vertex AI Integration
Full support for Google Cloud's Vertex AI platform.
- •Gemini 2 Pro & Gemini 2 Flash
- •PaLM 2 (legacy)
- •Imagen 2
- •Embeddings
Team Workspaces
Organize your AI spending by team with dedicated workspaces.
- •Separate dashboards per team
- •Team-specific budgets and alerts
- •Role-based access control
- •Cross-team comparison views
Slack Integration
Get alerts where your team already works.
- •Real-time alert notifications
- •Weekly digest summaries
- •Slash commands for quick lookups
- •Interactive budget approvals
Node.js SDK v1.0
The official TokenTra Node.js SDK is now generally available.
- •Zero-latency wrapper for OpenAI, Anthropic, Google, Azure
- •Automatic token counting and cost calculation
- •Custom attribution tags
- •TypeScript support
- •Batch telemetry (non-blocking)
Azure OpenAI Integration
Connect your Azure OpenAI deployments to TokenTra.
- •Service principal authentication
- •Cost data via Azure Cost Management API
- •Usage metrics via Azure Monitor
- •Per-deployment breakdown
Cost Attribution
Finally know who's spending what.
- •Attribution by Team, Project, Feature, User, and Custom tags
- •Chargeback reports for finance
- •Per-user unit economics
- •Feature cost analysis
TokenTra Launch 🚀
We're live! TokenTra launches with support for OpenAI and Anthropic.
- •Unified cost dashboard
- •Real-time sync (5-minute refresh)
- •Historical trends and comparison
- •Budget limits and alerts
- •Email notifications
- •CSV export
Subscribe to Updates
Get notified when we ship new features. No spam, just product updates.
Or follow us on Twitter/X for the latest updates.