February 2025 Summaries

Nitro and Floor Price shortcuts introduced by OpenRouter offer users efficient ways to prioritize either speed or cost when selecting providers for their AI models. The :nitro suffix sorts providers by throughput, ensuring the fastest option is chosen, while the :floor shortcut selects the lowest-cost provider, both acting as simplified equivalents to setting specific sorting criteria in the API. For users seeking more control over provider selection, advanced sorting options like "price," "throughput," or "latency" are available, overriding OpenRouter's default load balancing. Additionally, upcoming improvements will transition throughput metrics from averages to medians, promising more accurate real-world performance ratings and enhancing the overall speed and cost-effectiveness of AI access.

OpenRouter has introduced new shortcuts, Nitro and Floor Price, to optimize user experience by prioritizing speed and cost, respectively. The Nitro variant, indicated by the :nitro suffix, sorts providers by throughput to ensure users access the fastest available model, while the Floor Price shortcut, marked by the :floor suffix, selects the lowest-priced provider for cost-effective solutions. Users seeking more control can manually set sorting preferences by attributes such as price, throughput, or latency, overriding the default load balancing. Upcoming improvements include a shift from average to median throughput metrics, promising enhanced speed ratings that align more closely with real-world performance, ultimately facilitating faster and more economical AI access.

Cloudflare has become a new provider on OpenRouter by offering its Workers AI platform, which now supports various open-source models accessible with low latency and competitive pricing. Among the models available from Cloudflare are Gemma 7B-IT, DeepSeek R1, Distill Qwen 32B, and several versions of Llama models, including Llama 3.3 70B, Llama 3.1 70B, and others with varying parameters. This integration allows users to explore and utilize a range of Cloudflare-supported models through OpenRouter, enhancing the accessibility of AI models for developers and businesses.

Cloudflare has become a new provider on OpenRouter, making its Workers AI platform available for serving a variety of open-source models with low latency and competitive pricing. The models now accessible through Cloudflare on OpenRouter include Gemma 7B-IT with tool calling, DeepSeek R1, Distill Qwen 32B, Llama 3.3 70B, Llama 3.1 70B, Llama 3.2 11B Vision Instruct, Llama 3.1 8B, Llama 3 8B, Llama 3.2 3B, and Llama 3.2 1B. This partnership allows users to explore a wide range of Cloudflare-supported models, enhancing options for developers and businesses utilizing OpenRouter.