The AI arena is free today

Open Superagent

Qwen3 VL 235B A22B Thinking vs Qwen3 VL 4B Instruct

Qwen3 VL 235B A22B Thinking significantly outperforms across most benchmarks. Qwen3 VL 4B Instruct is 5.4x cheaper per token.

Alibaba Cloud / Qwen Team · Alibaba Cloud / Qwen Team · Updated for 2026

Which is better?

Qwen3 VL 235B A22B Thinking outperforms in 37 benchmarks (AI2D, AIME 2025, BFCL-v3, BLINK, CC-OCR, CharadesSTA, CharXiv-R, DocVQAtest, ERQA, Hallusion Bench, HMMT25, IFEval, Include, InfoVQAtest, LiveBench 20241125, LiveCodeBench v6, LVBench, MathVision, MathVista-Mini, MMBench-V1.1, MMLU, MMLU-Pro, MMLU-ProX, MMLU-Redux, MM-MT-Bench, MMMU-Pro, MMStar, MuirBench, OCRBench-V2 (en), OCRBench-V2 (zh), OSWorld, RealWorldQA, ScreenSpot, ScreenSpot Pro, SuperGPQA, VideoMMMU, WritingBench), while Qwen3 VL 4B Instruct is better at 3 benchmarks (OCRBench, ODinW, SimpleQA). Qwen3 VL 235B A22B Thinking significantly outperforms across most benchmarks.

On price, Qwen3 VL 4B Instruct is roughly 5.4x cheaper per token on a blended 3:1 input/output basis, which adds up quickly at production volume.

Based on current benchmark, pricing, and model metadata for 2026.

Choose Qwen3 VL 235B A22B Thinking

  • you want the strongest raw capability — it leads on 37 of 40 shared benchmarks

Choose Qwen3 VL 4B Instruct

  • cost matters — it's about 5.4x cheaper per token

At a glance

The differences that matter most.

Benchmark wins
37 of 40
3 of 40
Input price
$0.45 / M
$0.10 / M
Output price
$3.49 / M
$0.60 / M
Context window
262,144
262,144
Released
Sep 2025
Sep 2025
License
Apache 2.0
Apache 2.0

Performance Benchmarks

Comparative analysis across standard metrics

40 benchmarks

Qwen3 VL 235B A22B Thinking outperforms in 37 benchmarks (AI2D, AIME 2025, BFCL-v3, BLINK, CC-OCR, CharadesSTA, CharXiv-R, DocVQAtest, ERQA, Hallusion Bench, HMMT25, IFEval, Include, InfoVQAtest, LiveBench 20241125, LiveCodeBench v6, LVBench, MathVision, MathVista-Mini, MMBench-V1.1, MMLU, MMLU-Pro, MMLU-ProX, MMLU-Redux, MM-MT-Bench, MMMU-Pro, MMStar, MuirBench, OCRBench-V2 (en), OCRBench-V2 (zh), OSWorld, RealWorldQA, ScreenSpot, ScreenSpot Pro, SuperGPQA, VideoMMMU, WritingBench), while Qwen3 VL 4B Instruct is better at 3 benchmarks (OCRBench, ODinW, SimpleQA).

Qwen3 VL 235B A22B Thinking significantly outperforms across most benchmarks.

Tue Aug 25 2026 • llm-stats.com

Arena Performance

Playground indexes and blind preference scores

Pricing Analysis

Price comparison per million tokens

Qwen3 VL 4B Instruct costs less

For input processing, Qwen3 VL 235B A22B Thinking ($0.45/1M tokens) is 4.5x more expensive than Qwen3 VL 4B Instruct ($0.10/1M tokens).

For output processing, Qwen3 VL 235B A22B Thinking ($3.49/1M tokens) is 5.8x more expensive than Qwen3 VL 4B Instruct ($0.60/1M tokens).

In conclusion, Qwen3 VL 235B A22B Thinking is more expensive than Qwen3 VL 4B Instruct.*

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

Lowest available price from all providers
Tue Aug 25 2026 • llm-stats.com
Alibaba Cloud / Qwen Team
Qwen3 VL 235B A22B Thinking
Input tokens$0.45
Output tokens$3.49
Best providerDeepinfra
Alibaba Cloud / Qwen Team
Qwen3 VL 4B Instruct
Input tokens$0.10
Output tokens$0.60
Best providerDeepinfra
Notice missing or incorrect data?Start an Issue

Model Size

Parameter count comparison

232.0B diff

Qwen3 VL 235B A22B Thinking has 232.0B more parameters than Qwen3 VL 4B Instruct, making it 5800.0% larger.

Alibaba Cloud / Qwen Team
Qwen3 VL 235B A22B Thinking
236.0Bparameters
Alibaba Cloud / Qwen Team
Qwen3 VL 4B Instruct
4.0Bparameters
236.0B
Qwen3 VL 235B A22B Thinking
4.0B
Qwen3 VL 4B Instruct

Context Window

Maximum input and output token capacity

Both models have the same input context window of 262,144 tokens. Both models can generate responses up to 262,144 tokens.

Alibaba Cloud / Qwen Team
Qwen3 VL 235B A22B Thinking
Input262,144 tokens
Output262,144 tokens
Alibaba Cloud / Qwen Team
Qwen3 VL 4B Instruct
Input262,144 tokens
Output262,144 tokens
Tue Aug 25 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both Qwen3 VL 235B A22B Thinking and Qwen3 VL 4B Instruct support multimodal inputs.

They are both capable of processing various types of data, offering versatility in application.

Qwen3 VL 235B A22B Thinking

Text
Images
Audio
Video

Qwen3 VL 4B Instruct

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 VL 235B A22B Thinking

Apache 2.0

Open weights

Qwen3 VL 4B Instruct

Apache 2.0

Open weights

Release Timeline

When each model was launched

Both models were released on 2025-09-22.

They likely represent similar generations of model development.

Qwen3 VL 235B A22B Thinking

Sep 22, 2025

11 months ago

Qwen3 VL 4B Instruct

Sep 22, 2025

11 months ago

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 VL 235B A22B Thinking is available from DeepInfra, Novita. Qwen3 VL 4B Instruct is available from DeepInfra.

Qwen3 VL 235B A22B Thinking

deepinfra logo
Deepinfra
Input Price:Input: $0.45/1MOutput Price:Output: $3.49/1M
novita logo
Novita
Input Price:Input: $0.98/1MOutput Price:Output: $3.95/1M

Qwen3 VL 4B Instruct

deepinfra logo
Deepinfra
Input Price:Input: $0.10/1MOutput Price:Output: $0.60/1M
* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion

Judge for yourself.

Run your own prompts against Qwen3 VL 235B A22B Thinking and Qwen3 VL 4B Instruct side-by-side, then vote on the output you prefer.

Qwen3 VL 235B A22B Thinking
✓ Preferred
Qwen3 VL 4B Instruct
Open in Playground

FAQ

Common questions about Qwen3 VL 235B A22B Thinking vs Qwen3 VL 4B Instruct.

Which is better, Qwen3 VL 235B A22B Thinking or Qwen3 VL 4B Instruct?

Qwen3 VL 235B A22B Thinking significantly outperforms across most benchmarks. Qwen3 VL 235B A22B Thinking is made by Alibaba Cloud / Qwen Team and Qwen3 VL 4B Instruct 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 VL 235B A22B Thinking compare to Qwen3 VL 4B Instruct in benchmarks?

Qwen3 VL 235B A22B Thinking scores ZebraLogic: 97.3%, DocVQAtest: 96.5%, ScreenSpot: 95.4%, CountBench: 93.7%, MMLU-Redux: 93.7%. Qwen3 VL 4B Instruct scores DocVQAtest: 95.3%, ScreenSpot: 94.0%, OCRBench: 88.1%, MMBench-V1.1: 85.1%, AI2D: 84.1%.

Is Qwen3 VL 235B A22B Thinking cheaper than Qwen3 VL 4B Instruct?

Qwen3 VL 4B Instruct is 4.5x cheaper for input tokens. Qwen3 VL 235B A22B Thinking costs $0.45/M input and $3.49/M output via deepinfra. Qwen3 VL 4B Instruct costs $0.10/M input and $0.60/M output via deepinfra.

What are the context window sizes for Qwen3 VL 235B A22B Thinking and Qwen3 VL 4B Instruct?

Qwen3 VL 235B A22B Thinking supports 262K tokens and Qwen3 VL 4B Instruct supports 262K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Qwen3 VL 235B A22B Thinking and Qwen3 VL 4B Instruct?

Key differences include input pricing ($0.45 vs $0.10/M). See the full comparison above for benchmark-by-benchmark results.