B01-NUna Documentation
An intelligent AI orchestration system with 280+ billion parameters across specialized models. Features GPU-accelerated self-learning, intelligent multi-model routing, and real-time inference optimization.
Model Cards
B01-NUna Model Card
Complete orchestration system (280B+ parameters) with intelligent routing and GPU-accelerated self-learning.
View Model CardB01-1.2V-5B Model Card
Foundation model component (5.2B parameters) within the B01-NUna orchestration system.
Download PDFNote: B01-1.2V-5B (5.2B parameters) is the foundation model component within the B01-NUna orchestration system.
Quickstart
Get started with the B01-NUna API in seconds. We support two authentication methods:
Method 1: Authorization Header
curl -X POST https://helloblue.ai/api/v1/generate \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "b01-nuna",
"prompt": "Explain quantum computing in simple terms.",
"max_tokens": 512
}'Method 2: X-API-Key Header
curl -X POST https://helloblue.ai/api/v1/generate \
-H "X-API-Key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "b01-nuna",
"prompt": "Explain quantum computing in simple terms.",
"max_tokens": 512
}'Replace YOUR_API_KEY with your actual API key. Both authentication methods are fully supported.
Supported Models: b01-1.2v-5b (Foundation Model), b01-nuna (Ultra-Fast Inference)
API Reference
Endpoint Details
- Endpoint:
POST https://helloblue.ai/api/v1/generate - Models:
b01-1.2v-5b(Foundation),b01-nuna(Ultra-Fast) - Auth:Bearer token or X-API-Key header
- Content-Type:
application/json
Request Parameters
prompt(string, required) - The input textmax_tokens(integer, optional) - Maximum tokens to generatetemperature(float, optional) - Response creativity (0.0-1.0)model(string, required) - Model identifier
Response Format
{
"model": "b01-nuna",
"prompt": "Explain quantum computing in simple terms.",
"max_tokens": 512,
"text": "Of course! Quantum computing is a new way...",
"object": "text_completion",
"usage": {
"prompt_tokens": 12,
"completion_tokens": 366,
"total_tokens": 378
},
"metadata": {
"provider": "Enterprise AI Service",
"processing_time": 5810,
"confidence": 0.85,
"endpoint": "https://helloblue.ai/api/v1/generate",
"retries": 0,
"fallbacks": [],
"model_used": "b01-nuna"
}
}API Testing
Test our API endpoints with these ready-to-use curl commands. No authentication required for test endpoints.
System Health Check
Get comprehensive system health metrics, memory usage, and performance data.
curl -X GET https://helloblue.ai/api/health \ -H "Content-Type: application/json"
Simple AI Test (Direct Ollama)
Test direct communication with the local Ollama instance.
curl -X POST https://helloblue.ai/api/test-simple \
-H "Content-Type: application/json" \
-d '{
"messages": [
{"role": "user", "content": "What is artificial intelligence?"}
]
}'Enterprise AI Test
Test the full enterprise AI service with fallback capabilities.
curl -X POST https://helloblue.ai/api/test-ai \
-H "Content-Type: application/json" \
-d '{
"messages": [
{"role": "system", "content": "You are a helpful AI assistant."},
{"role": "user", "content": "Explain quantum computing in simple terms."}
]
}'Testing Tips:
- All endpoints use
https://helloblue.aias the base URL - Test endpoints don't require API keys for basic functionality
- Use the health endpoints to monitor system status
- Production API endpoint requires authentication (Bearer token or X-API-Key)
Integration Features
Authentication & Security
- Dual authentication methods (Bearer token & X-API-Key)
- API key validation and rate limiting
- Secure header handling with middleware
AI Orchestration
- Enterprise AI Service integration
- Ollama fallback system
- Intelligent model routing
Real-time Features
- Live performance monitoring
- Memory optimization
- Health check endpoints
Model Benchmarks
Figures below are Meta's published evaluations for Llama 3.3 70B Instruct (default cloud backbone via Groq). They are not Helloblue-measured percentiles. See the model card for full context.
| Benchmark | Score | Protocol (Meta) |
|---|---|---|
| MMLU (CoT) | 86.0% | 0-shot, macro_avg / acc |
| MMLU Pro (CoT) | 68.9% | 5-shot, macro_avg / acc |
| IFEval | 92.1% | Steerability (per Meta table) |
| GPQA Diamond (CoT) | 50.5% | 0-shot, acc |
| HumanEval | 88.4% | 0-shot, pass@1 |
| MBPP EvalPlus (base) | 87.6% | 0-shot, pass@1 |
| MATH (CoT) | 77.0% | 0-shot, sympy_intersection_score |
| MGSM | 91.1% | 0-shot, exact match (multilingual grade-school math) |
| BFCL v2 | 77.3% | 0-shot, overall_ast_summary / macro_avg / valid |
Source: Meta Llama 3.3 MODEL_CARD.md