AI Models / Compare

Qwen3 235B-A22B

Best Qwen3 model — 235B total MoE, competitive with o1 on reasoning benchmarks.

Creator: Qwen
Lifecycle: Active
Context: 262.1K
Max output: 32.8K
Released: Apr 29, 2025
Status: unknown

Input: $0.70 / 1M tokens
Output: $2.80 / 1M tokens
Cached read: — / 1M tokens
Cached write: — / 1M tokens
Batch discount: —%
Source: Qwen3 235B-A22B pricing
Verified: Apr 5, 2026 (High)

Capabilities

Modalities: text→text
Capabilities: reasoningbatchSupportpromptCachingfunctionCallingstructuredOutputs
Strengths: Open frontier quality, 119 languages, Apache 2.0
Tradeoffs: Text-only, needs multi-GPU to self-host

Official Links

Intro

https://qwenlm.github.io/blog/qwen3/

Docs

https://qwenlm.github.io/blog/

Pricing

https://www.alibabacloud.com/help/en/model-studio/model-pricing

Release note

https://qwenlm.github.io/blog/

Benchmark Coverage

Benchmark	Version	Score	Date	Source	Notes
AIME 2024	2024	85.7 %	Apr 1, 2025	Qwen Team	Thinking mode, vendor-reported

Release History

Release	Alias	Lifecycle	Release Date	Deprecation	Shutdown	Summary
Qwen3 235B-A22B	qwen3-235b-a22b	Active	Apr 29, 2025	—	—	Current published model family snapshot.

Host Coverage

Host	Type	Context	Pricing Note	Differences
Alibaba DashScope	first-party	262.1K	$0.70/$2.80 per MTok (standard tier).	Hybrid thinking mode

Migration Guidance

Best open MoE reasoning model. Use Qwen3-32B for single-GPU deployments.

Change Events

Date	Type	Title	Description	Source
Apr 29, 2025	family_added	Qwen3 235B-A22B published	Initial public model family launch.	Qwen3 235B-A22B release notes

Other models from Qwen

Qwen2.5 72B, Qwen3 32B, QwQ-32B