Qwen2.5 Coder 32B Instruct

Text input Text output Free Option
Author's Description

Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen). Qwen2.5-Coder brings the following improvements upon CodeQwen1.5: - Significantly improvements in **code generation**, **code reasoning** and **code fixing**. - A more comprehensive foundation for real-world applications such as **Code Agents**. Not only enhancing coding capabilities but also maintaining its strengths in mathematics and general competencies. To read more about its evaluation results, check out [Qwen 2.5 Coder's blog](https://qwenlm.github.io/blog/qwen2.5-coder-family/).

Key Specifications
Cost
$
Context
32K
Parameters
500B (Rumoured)
Released
Nov 11, 2024
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Response Format Stop Seed Min P Top P Max Tokens Frequency Penalty Temperature Presence Penalty
Features

This model supports the following features:

Response Format
Performance Summary

Qwen2.5 Coder 32B Instruct, released on November 11, 2024, demonstrates a strong overall performance profile, particularly excelling in its core domain of coding. The model generally performs among the fastest models, ranking in the 73rd percentile for speed across eight benchmarks. It consistently offers highly competitive pricing, placing in the 87th percentile across seven benchmarks. Notably, its reliability is exceptional, achieving a 100% success rate across all eight benchmarks, indicating minimal technical failures. In terms of specific benchmark results, Qwen2.5 Coder 32B Instruct achieved perfect accuracy in Hallucinations (Baseline) and one of the Instruction Following (Baseline) tests, highlighting its ability to acknowledge uncertainty and precisely follow directives. It also showed strong performance in Ethics (99.0% accuracy) and Reasoning (76.0% accuracy). While its General Knowledge (96.2%) and Email Classification (96.0%) scores are solid, they are not top-tier. The model's Coding performance, at 77.0% accuracy, is respectable but falls in the 39th percentile, suggesting room for improvement despite its "Coder" designation. A notable weakness is the second Instruction Following benchmark, where it scored 54.5% accuracy, indicating inconsistency in handling certain types of complex instructions.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.06
Completion $0.15

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
DeepInfra
DeepInfra | qwen/qwen-2.5-coder-32b-instruct 32K $0.06 / 1M tokens $0.15 / 1M tokens
Nebius
Nebius | qwen/qwen-2.5-coder-32b-instruct 131K $0.04 / 1M tokens $0.14 / 1M tokens
Lambda
Lambda | qwen/qwen-2.5-coder-32b-instruct 32K $0.04 / 1M tokens $0.14 / 1M tokens
Hyperbolic
Hyperbolic | qwen/qwen-2.5-coder-32b-instruct 32K $0.2 / 1M tokens $0.2 / 1M tokens
Cloudflare
Cloudflare | qwen/qwen-2.5-coder-32b-instruct 32K $0.66 / 1M tokens $1 / 1M tokens
Together
Together | qwen/qwen-2.5-coder-32b-instruct 16K $0.8 / 1M tokens $0.8 / 1M tokens
Featherless
Featherless | qwen/qwen-2.5-coder-32b-instruct 16K $0.04 / 1M tokens $0.14 / 1M tokens
Chutes
Chutes | qwen/qwen-2.5-coder-32b-instruct 32K $0.04 / 1M tokens $0.14 / 1M tokens
Chutes
Chutes | qwen/qwen-2.5-coder-32b-instruct 32K $0.04 / 1M tokens $0.14 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by qwen