DeepSeek: R1 Distill Llama 70B

DeepSeek R1 Distill Llama 70B is an advanced large language model, meticulously distilled from Llama-3.3-70B-Instruct and fine-tuned using outputs from DeepSeek R1. This innovative approach allows it to achieve exceptional performance across various benchmarks, including AIME 2024 pass@1: 70.0, MATH-500 pass@1: 94.5, and a CodeForces Rating of 1633. Designed for efficiency and power, this model excels in complex tasks requiring strong reasoning, mathematical problem-solving, and code generation. It supports a generous context window of 131K tokens and can generate up to 4K tokens in output. Capabilities include functions, code interpretation, and streaming responses. Access DeepSeek R1 Distill Llama 70B for free on Multi AI. Pricing is highly competitive at $0.03 per 1M input tokens and $0.11 per 1M output tokens, making it an ideal choice for high-performance, cost-effective AI applications.