CompactifAI API
#35 v OstatníMultiverse Computing · od 11. Juni 2025 · 2× · naposledy 28. 7. 2026
14
Momentum
CompactifAI API by Multiverse Computing is a cloud inference service providing access to compressed AI language models (including Llama, Mistral, DeepSeek). It offers an OpenAI-compatible interface with chat and text completions, multimodal support, and streaming. Pricing is usage-based (pay-as-you-go) per token processed, and the API is available via AWS Marketplace among other channels. The product aims to significantly reduce compute costs, energy consumption, and latency compared to uncompressed models through model compression.
Vývoj momenta
16.05.14.08.
Vlastnosti
| License | Commercial API access with clearly defined licensing terms per model |
| Platform | Cloud API (OpenAI-compatible), available via AWS Marketplace among others (AWS SageMaker Hyperpod) |
| Price | Pay-as-you-go, e.g. Hypernova 60B: $0.04/M input, $0.14/M output tokens |
| Release Date | June 11, 2025 (Initial release per changelog) |