Model Registry & Latency Benchmarks
Catalog of connected foundation models, prompt caching support, and evaluated empirical latency.
Gemini 2.0 Flash
Context Window: 1,048,576 tokens
MultimodalReasoningFast Ingestion
In: $0.10 / 1MOut: $0.40 / 1MAccuracy: 94.2%
Claude 3.5 Sonnet
Context Window: 200,000 tokens
Tool CallingFaithful Explanations
In: $3.00 / 1MOut: $15.00 / 1MAccuracy: 96.8%
Llama 3.1 70B (Ollama)
Context Window: 128,000 tokens
Air-GappedZero Cloud Egress
In: $0.00 (Self-hosted)Out: $0.00 (Self-hosted)Accuracy: 91.4%
DeepSeek V3
Context Window: 64,000 tokens
Cost-EfficientCode & Math
In: $0.14 / 1MOut: $0.28 / 1MAccuracy: 92.6%