Qwen: Qwen3 235B A22B

Text input Text output
Author's Description

Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass. It supports seamless switching between a "thinking" mode for complex reasoning, math, and code tasks, and a "non-thinking" mode for general conversational efficiency. The model demonstrates strong reasoning ability, multilingual support (100+ languages and dialects), advanced instruction-following, and agent tool-calling capabilities. It natively handles a 32K token context window and extends up to 131K tokens using YaRN-based scaling.

Key Specifications
Cost
$$$$
Context
40K
Parameters
235B
Released
Apr 28, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Stop Presence Penalty Temperature Seed Response Format Frequency Penalty Max Tokens Include Reasoning Tool Choice Top P Min P Tools Reasoning
Features

This model supports the following features:

Tools Reasoning Response Format
Performance Summary

Qwen3-235B-A22B, a 235B parameter MoE model, demonstrates exceptional reliability, consistently providing usable responses with minimal technical failures, ranking in the 100th percentile. However, its speed performance is a notable area for improvement, with response times tending to be longer, placing it in the 7th percentile across benchmarks. Pricing is moderate, falling within the 36th percentile. The model exhibits outstanding accuracy across several critical benchmarks. It achieved perfect scores in Email Classification, Reasoning, and General Knowledge, often being the most accurate model at its price point and among models of similar speed. Its "thinking" mode appears highly effective for complex tasks. While its Ethics performance is strong at 99.0% accuracy, its Instruction Following accuracy is comparatively lower at 40.4%, suggesting potential limitations in handling highly complex, multi-layered instructions. Its ability to seamlessly switch between "thinking" and "non-thinking" modes, coupled with strong multilingual support and agent tool-calling capabilities, positions it as a versatile model, despite its slower processing times.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.13
Completion $0.6

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
DeepInfra
DeepInfra | qwen/qwen3-235b-a22b-04-28 40K $0.13 / 1M tokens $0.6 / 1M tokens
Parasail
Parasail | qwen/qwen3-235b-a22b-04-28 40K $0.13 / 1M tokens $0.6 / 1M tokens
Kluster
Kluster | qwen/qwen3-235b-a22b-04-28 40K $0.13 / 1M tokens $0.6 / 1M tokens
GMICloud
GMICloud | qwen/qwen3-235b-a22b-04-28 32K $0.13 / 1M tokens $0.6 / 1M tokens
Together
Together | qwen/qwen3-235b-a22b-04-28 40K $0.2 / 1M tokens $0.6 / 1M tokens
Nebius
Nebius | qwen/qwen3-235b-a22b-04-28 40K $0.2 / 1M tokens $0.6 / 1M tokens
Novita
Novita | qwen/qwen3-235b-a22b-04-28 40K $0.13 / 1M tokens $0.6 / 1M tokens
Fireworks
Fireworks | qwen/qwen3-235b-a22b-04-28 131K $0.22 / 1M tokens $0.88 / 1M tokens
Friendli
Friendli | qwen/qwen3-235b-a22b-04-28 131K $0.13 / 1M tokens $0.6 / 1M tokens
Cerebras
Cerebras | qwen/qwen3-235b-a22b-04-28 40K $0.13 / 1M tokens $0.6 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Free Executions Accuracy Cost Duration
Other Models by qwen