Model Comparison

Qwen3 235B A22B vs Qwen3 VL 30B A3B Thinking

Qwen3 VL 30B A3B Thinking significantly outperforms across most benchmarks. Qwen3 235B A22B is 4.0x cheaper per token.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

7 benchmarks

Qwen3 235B A22B outperforms in 1 benchmarks (MMLU), while Qwen3 VL 30B A3B Thinking is better at 6 benchmarks (AIME 2025, GPQA, Include, MMLU-Pro, MMLU-Redux, SuperGPQA).

Qwen3 VL 30B A3B Thinking significantly outperforms across most benchmarks.

Tue May 12 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Qwen3 235B A22B costs less

For input processing, Qwen3 235B A22B ($0.10/1M tokens) is 2.0x cheaper than Qwen3 VL 30B A3B Thinking ($0.20/1M tokens).

For output processing, Qwen3 235B A22B ($0.10/1M tokens) is 9.9x cheaper than Qwen3 VL 30B A3B Thinking ($0.99/1M tokens).

In conclusion, Qwen3 VL 30B A3B Thinking is more expensive than Qwen3 235B A22B.*

* Using a 3:1 ratio of input to output tokens

Lowest available price from all providers

Tue May 12 2026 • llm-stats.com

Qwen3 235B A22B

Input tokens$0.10

Output tokens$0.10

Best providerFireworks

Qwen3 VL 30B A3B Thinking

Input tokens$0.20

Output tokens$0.99

Best providerNovita

Notice missing or incorrect data?Start an Issue→

Model Size

Parameter count comparison

204.0B diff

Qwen3 235B A22B has 204.0B more parameters than Qwen3 VL 30B A3B Thinking, making it 658.1% larger.

Qwen3 235B A22B

235.0Bparameters

Qwen3 VL 30B A3B Thinking

31.0Bparameters

235.0B

Qwen3 235B A22B

31.0B

Qwen3 VL 30B A3B Thinking

Context Window

Maximum input and output token capacity

Qwen3 VL 30B A3B Thinking accepts 131,072 input tokens compared to Qwen3 235B A22B's 128,000 tokens. Qwen3 235B A22B can generate longer responses up to 128,000 tokens, while Qwen3 VL 30B A3B Thinking is limited to 32,768 tokens.

Qwen3 235B A22B

Input128,000 tokens

Output128,000 tokens

Qwen3 VL 30B A3B Thinking

Input131,072 tokens

Output32,768 tokens

Tue May 12 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Qwen3 VL 30B A3B Thinking supports multimodal inputs, whereas Qwen3 235B A22B does not.

Qwen3 VL 30B A3B Thinking can handle both text and other forms of data like images, making it suitable for multimodal applications.

Qwen3 235B A22B

Text

Images

Audio

Video

Qwen3 VL 30B A3B Thinking

Text

Images

Audio

Video

License

Usage and distribution terms

Both models are licensed under Apache 2.0.

Both models share the same licensing terms, providing consistent usage rights.

Qwen3 235B A22B

Apache 2.0

Open weights

Qwen3 VL 30B A3B Thinking

Apache 2.0

Open weights

Release Timeline

When each model was launched

Qwen3 235B A22B was released on 2025-04-29, while Qwen3 VL 30B A3B Thinking was released on 2025-09-22.

Qwen3 VL 30B A3B Thinking is 5 months newer than Qwen3 235B A22B.

Qwen3 235B A22B

Apr 29, 2025

1.0 years ago

Qwen3 VL 30B A3B Thinking

Sep 22, 2025

7 months ago

4mo newer

Knowledge Cutoff

When training data ends

Neither model specifies a knowledge cutoff date.

Unable to compare the recency of their training data.

No cutoff dates available

Provider Availability

Qwen3 235B A22B is available from Fireworks, DeepInfra, Novita, Together. Qwen3 VL 30B A3B Thinking is available from Novita, DeepInfra.

Qwen3 235B A22B

Fireworks

Input Price:Input: $0.10/1MOutput Price:Output: $0.10/1M

Deepinfra

Input Price:Input: $0.20/1MOutput Price:Output: $0.60/1M

Novita

Input Price:Input: $0.20/1MOutput Price:Output: $0.80/1M

Together

Input Price:Input: $0.20/1MOutput Price:Output: $0.60/1M

Qwen3 VL 30B A3B Thinking

Novita

Input Price:Input: $0.20/1MOutput Price:Output: $1.00/1M

Deepinfra

Input Price:Input: $0.29/1MOutput Price:Output: $0.99/1M

* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Qwen3 235B A22B

View details

Alibaba Cloud / Qwen Team

Less expensive input tokens

Less expensive output tokens

Higher MMLU score (87.8% vs 87.6%)

Qwen3 VL 30B A3B Thinking

View details

Alibaba Cloud / Qwen Team

Larger context window (131,072 tokens)

Supports multimodal inputs

Higher AIME 2025 score (83.1% vs 81.5%)

Higher GPQA score (74.4% vs 47.5%)

Higher Include score (74.5% vs 73.5%)

Higher MMLU-Pro score (80.5% vs 68.2%)

Higher MMLU-Redux score (90.9% vs 87.4%)

Higher SuperGPQA score (56.4% vs 44.1%)

Detailed Comparison

AI Model Comparison Table
Feature	Qwen3 235B A22B	Qwen3 VL 30B A3B Thinking

FAQ

Common questions about Qwen3 235B A22B vs Qwen3 VL 30B A3B Thinking.

Which is better, Qwen3 235B A22B or Qwen3 VL 30B A3B Thinking?

Qwen3 VL 30B A3B Thinking significantly outperforms across most benchmarks. Qwen3 235B A22B is made by Alibaba Cloud / Qwen Team and Qwen3 VL 30B A3B Thinking is made by Alibaba Cloud / Qwen Team. The best choice depends on your use case — compare their benchmark scores, pricing, and capabilities above.

How does Qwen3 235B A22B compare to Qwen3 VL 30B A3B Thinking in benchmarks?

Qwen3 235B A22B scores Arena Hard: 95.6%, GSM8k: 94.4%, BBH: 88.9%, MMLU: 87.8%, MMLU-Redux: 87.4%. Qwen3 VL 30B A3B Thinking scores DocVQAtest: 95.0%, ScreenSpot: 94.7%, MMLU-Redux: 90.9%, MMBench-V1.1: 88.9%, MMLU: 87.6%.

Is Qwen3 235B A22B cheaper than Qwen3 VL 30B A3B Thinking?

Qwen3 235B A22B is 2.0x cheaper for input tokens. Qwen3 235B A22B costs $0.10/M input and $0.10/M output via fireworks. Qwen3 VL 30B A3B Thinking costs $0.20/M input and $0.99/M output via novita.

What are the context window sizes for Qwen3 235B A22B and Qwen3 VL 30B A3B Thinking?

Qwen3 235B A22B supports 128K tokens and Qwen3 VL 30B A3B Thinking supports 131K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Qwen3 235B A22B and Qwen3 VL 30B A3B Thinking?

Key differences include context window (128K vs 131K), input pricing ($0.10 vs $0.20/M), multimodal support (no vs yes). See the full comparison above for benchmark-by-benchmark results.