GPT-6.1 Terra1.5x
gpt-6.1-terraVisual reasoning, natural writing and general-purpose tasks.
VANGU / COMPANY
Dependable AI Infrastructure for Engineers
We engineer predictable, transparent and resilient API infrastructure—free from arbitrary lock-ins, opaque quotas, or surprise billing.
%99.98
Annual Uptime Target
Monitoring, resource limits and recovery controls support service availability.
Time-limited
Content Retention
Conversation content and technical records follow configured retention periods.
SSE
Streaming Responses
Receive text as it is generated on supported streaming endpoints.
00.00
Automatic Quota Renewal
Daily allowances cleanly refresh every midnight without rollover friction.
OUR MISSION
The Next Benchmark in AI Integration
While foundational models evolve at breakneck speed, engineering teams and production services struggle with two major friction points: volatile, unpredictable usage bills and opaque quota architectures.
Vangu was built to make AI consumption as predictable, stable and transparent as drawing power from a grid. We support your projects with open API conventions, clear usage information and time-limited content retention.
ENGINEERING PHILOSOPHY
Engineered For Developers, By Developers
A gateway is valuable for its developer experience, documentation and usage visibility. Vangu supports OpenAI and Anthropic request formats through a shared API.
Choose an entitled model from the catalog and connect using OpenAI-compatible Chat, native Responses or Anthropic Messages. The platform manages provider selection and usage records.
Zero Overages Guarantee
Depleted limits yield clean, predictable halts without surprise billing or stealth credit card debits. Engineering leaders retain full budget control.
Encrypted, Time-limited Content Retention
When content retention is enabled, conversation history is stored encrypted for the configured period. Usage, cache and request replay records follow their own retention periods.
Transparent HTTP Telemetry
Use request IDs, status codes and applicable remaining-quota/reset headers to trace calls. Inspect model, token, cache and latency records in your dashboard.
Streaming Responses with SSE
Receive generated text incrementally on supported endpoints. First-token and total response times vary by model, context length and provider load.
MODEL ARCHITECTURE
Frontier Model Family & Dynamic Multipliers
Dynamically route across specialized performance tiers with a single key and unified quota pool.
GPT-6.1 Sol2x
gpt-6.1-solAdvanced coding, debugging, architecture analysis and multi-step reasoning.
GPT-6 Astra4x
gpt-6-astraComplex analysis, research and tasks requiring deep reasoning.
GPT-6 Sol2x
gpt-6-solCoding, tool-using agents and multi-step technical tasks.
GPT-6 Luna1x
gpt-6-lunaEveryday chat, summarization, classification and high-volume text tasks.
GPT-5.6 Sol2x
gpt-5.6-solCode generation, structured analysis and compatibility with existing workflows.
GPT-5.6 Terra1.5x
gpt-5.6-terraGeneral text generation, translation and balanced everyday tasks.
GPT-5.6 Luna1x
gpt-5.6-lunaShort answers, simple transformations and lightweight assistant tasks.
GPT Image 2Image
gpt-image-2Visual drafts, prototypes and standard image generation.
GPT Image 2.5Image
gpt-image-2.5Image generation and editing with selectable Flare or Sunburst modes.
GPT Image 2.5 FlareImage
gpt-image-2.5-flareA model ID providing direct access to the Flare image mode.
GPT Image 2.5 SunburstImage
gpt-image-2.5-sunburstA model ID providing direct access to the Sunburst image mode.
This list is refreshed from the live model catalog. Multipliers are general catalog information; your purchased plan or key billing mode determines access, pricing and allowances. Image models use separate image allowances or tariffs.
SECURITY STANDARDS
Enterprise-Grade Security & Compliance
256-Bit TLS 1.3 Encryption
All gateway traffic is strictly encrypted in transit using TLS 1.3 and HSTS, enforcing modern cipher suites across all endpoints.
PCI-DSS Compliant Payments
No raw payment credentials touch our servers. Subscriptions are tokenized and processed via PCI-DSS Level 1 compliant infrastructure.
Granular Keys & 2FA
Two-factor authentication (TOTP) safeguards your account. Generate scoped API keys with dedicated usage caps for separate environments.
FREQUENTLY ASKED
Corporate & Platform Inquiries
Are my prompts or outputs used for model training?
Vangu does not use your prompts or responses for model training. Requests are sent to the selected provider for inference, and its data processing terms also apply. Conversation history and technical records follow configured retention periods.
How long does migration to Vangu take?
For OpenAI-compatible clients, configure https://api.vangu.dev/v1, supply your Vangu key and select an entitled model ID. Follow the app-specific documentation for Codex and Claude Code.
Do you provide official tax invoices for corporations?
Review payment and invoice records in your dashboard. Use payment details or support for company billing information and currently available documents.
What happens when my daily quota is exhausted?
The gateway safely returns a 429 response detailing the 00:00 UTC+3 reset time. No surprise overage fees or stealth charges will ever occur.