DeepSeek: R1 Distill Qwen 14B

Text input Text output Free Option
Author's Description

DeepSeek R1 Distill Qwen 14B is a distilled large language model based on [Qwen 2.5 14B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-14B), using outputs from [DeepSeek R1](/deepseek/deepseek-r1). It outperforms OpenAI's o1-mini across various benchmarks, achieving new state-of-the-art results for dense models. Other benchmark results include: - AIME 2024 pass@1: 69.7 - MATH-500 pass@1: 93.9 - CodeForces Rating: 1481 The model leverages fine-tuning from DeepSeek R1's outputs, enabling competitive performance comparable to larger frontier models.

Key Specifications
Cost
$$$
Context
64K
Parameters
14B
Released
Jan 29, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Seed Max Tokens Presence Penalty Frequency Penalty Logit Bias Include Reasoning Temperature Top P Stop Min P Reasoning
Features

This model supports the following features:

Reasoning
Performance Summary

DeepSeek R1 Distill Qwen 14B is a highly capable distilled large language model, demonstrating strong performance across various benchmarks. While it tends to exhibit longer response times, ranking in the 9th percentile for speed, it generally offers cost-effective solutions, placing in the 62nd percentile for price. The model showcases exceptional proficiency in Code Generation, achieving 93.0% accuracy in the Coding (Baseline) benchmark, positioning it in the 94th percentile and notably as the most accurate model at its price point. Its Reasoning capabilities are also robust, with an 86.0% accuracy, placing it in the 85th percentile. However, its performance in Ethics (87.5% accuracy), Email Classification (93.0% accuracy), and General Knowledge (77.5% accuracy) falls into the lower quartiles (25th-26th percentile), indicating areas for potential improvement. Overall, DeepSeek R1 Distill Qwen 14B's key strengths lie in its coding and reasoning abilities, making it a strong contender for tasks requiring complex problem-solving and code generation. Its primary weakness is its slower processing speed, which might impact real-time applications. Despite this, its competitive pricing and high accuracy in specific domains make it a valuable option, particularly where cost-efficiency and specialized performance are prioritized over raw speed.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.15
Completion $0.15

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Novita
Novita | deepseek/deepseek-r1-distill-qwen-14b 64K $0.15 / 1M tokens $0.15 / 1M tokens
GMICloud
GMICloud | deepseek/deepseek-r1-distill-qwen-14b 131K $0.15 / 1M tokens $0.15 / 1M tokens
Together
Together | deepseek/deepseek-r1-distill-qwen-14b 131K $0.15 / 1M tokens $0.15 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Free Executions Accuracy Cost Duration
Other Models by deepseek