Qwen3 VL 30B A3B Instruct vs Qwen3 VL 8B Thinking Comparison

Comparing Qwen3 VL 30B A3B Instruct and Qwen3 VL 8B Thinking across benchmarks, pricing, and capabilities.

Want to compare interactively?Try the playground

Performance Benchmarks

Comparative analysis across standard metrics

50 benchmarks

Qwen3 VL 30B A3B Instruct outperforms in 23 benchmarks (AI2D, Arena-Hard v2, BFCL-v3, CC-OCR, CharadesSTA, Creative Writing v3, GPQA, IFEval, Include, LVBench, MLVU-M, MMLU-Pro, MMLU-ProX, MM-MT-Bench, MMMU (val), MVBench, OCRBench, ODinW, RealWorldQA, ScreenSpot, ScreenSpot Pro, SuperGPQA, Video-MME), while Qwen3 VL 8B Thinking is better at 26 benchmarks (AIME 2025, BLINK, CharXiv-D, CharXiv-R, DocVQAtest, ERQA, Hallusion Bench, HMMT25, InfoVQAtest, LiveBench 20241125, LiveCodeBench v6, MathVision, MathVista-Mini, MMBench-V1.1, MMLU, MMLU-Redux, MMStar, MuirBench, Multi-IF, OCRBench-V2 (en), OCRBench-V2 (zh), OSWorld, PolyMATH, SimpleQA, VideoMMMU, WritingBench).

Qwen3 VL 8B Thinking has a slight edge in benchmark performance.

Sun Mar 15 2026 • llm-stats.com

Arena Performance

Human preference votes

Pricing Analysis

Price comparison per million tokens

Qwen3 VL 30B A3B Instruct costs less

For input processing, Qwen3 VL 30B A3B Instruct ($0.20/1M tokens) is 1.1x more expensive than Qwen3 VL 8B Thinking ($0.18/1M tokens).

For output processing, Qwen3 VL 30B A3B Instruct ($0.70/1M tokens) is 3.0x cheaper than Qwen3 VL 8B Thinking ($2.09/1M tokens).

In conclusion, Qwen3 VL 8B Thinking is more expensive than Qwen3 VL 30B A3B Instruct.*

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

Lowest available price from all providers

Sun Mar 15 2026 • llm-stats.com

Qwen3 VL 30B A3B Instruct

Input tokens$0.20

Output tokens$0.70

Best providerNovita

Qwen3 VL 8B Thinking

Input tokens$0.18

Output tokens$2.09

Best providerDeepinfra

Notice missing or incorrect data?Start an Issue→

Model Size

Parameter count comparison

22.0B diff

Qwen3 VL 30B A3B Instruct has 22.0B more parameters than Qwen3 VL 8B Thinking, making it 244.4% larger.

Qwen3 VL 30B A3B Instruct

31.0Bparameters

Qwen3 VL 8B Thinking

9.0Bparameters

31.0B

Qwen3 VL 30B A3B Instruct

9.0B

Qwen3 VL 8B Thinking

Context Window

Maximum input and output token capacity

Qwen3 VL 8B Thinking accepts 262,144 input tokens compared to Qwen3 VL 30B A3B Instruct's 131,072 tokens. Qwen3 VL 8B Thinking can generate longer responses up to 262,144 tokens, while Qwen3 VL 30B A3B Instruct is limited to 32,768 tokens.

Qwen3 VL 30B A3B Instruct

Input131,072 tokens

Output32,768 tokens

Qwen3 VL 8B Thinking

Input262,144 tokens

Output262,144 tokens

Sun Mar 15 2026 • llm-stats.com

Input Capabilities

Supported data types and modalities

Both Qwen3 VL 30B A3B Instruct and Qwen3 VL 8B Thinking support multimodal inputs.

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

Qwen3 VL 30B A3B Instruct

Text

Images

Audio

Video

Qwen3 VL 8B 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 VL 30B A3B Instruct

Apache 2.0

Open weights

Qwen3 VL 8B Thinking

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 30B A3B Instruct

Sep 22, 2025

5 months ago

Qwen3 VL 8B Thinking

Sep 22, 2025

5 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 30B A3B Instruct is available from Novita, DeepInfra. Qwen3 VL 8B Thinking is available from DeepInfra. The availability of providers can affect quality of the model and reliability.

Qwen3 VL 30B A3B Instruct

Novita

Input Price:Input: $0.20/1MOutput Price:Output: $0.70/1M

Deepinfra

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

Qwen3 VL 8B Thinking

Deepinfra

Input Price:Input: $0.18/1MOutput Price:Output: $2.09/1M

* Prices shown are per million tokens

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Key Takeaways

Qwen3 VL 30B A3B Instruct

View details

Alibaba Cloud / Qwen Team

Less expensive output tokens

Higher AI2D score (85.0% vs 84.9%)

Higher Arena-Hard v2 score (58.5% vs 51.1%)

Higher BFCL-v3 score (66.3% vs 63.0%)

Higher CC-OCR score (80.7% vs 76.3%)

Higher CharadesSTA score (63.5% vs 59.9%)

Higher Creative Writing v3 score (84.6% vs 82.4%)

Higher GPQA score (70.4% vs 69.9%)

Higher IFEval score (85.8% vs 83.2%)

Higher Include score (71.6% vs 69.5%)

Higher LVBench score (62.5% vs 55.8%)

Higher MLVU-M score (81.3% vs 75.1%)

Higher MMLU-Pro score (77.8% vs 77.3%)

Higher MMLU-ProX score (70.9% vs 70.7%)

Higher MM-MT-Bench score (8.1% vs 8.0%)

Higher MMMU (val) score (74.2% vs 74.1%)

Higher MVBench score (72.3% vs 69.0%)

Higher OCRBench score (90.3% vs 81.9%)

Higher ODinW score (47.5% vs 39.8%)

Higher RealWorldQA score (73.7% vs 73.5%)

Higher ScreenSpot score (94.7% vs 93.6%)

Higher ScreenSpot Pro score (60.5% vs 46.6%)

Higher SuperGPQA score (53.1% vs 51.2%)

Higher Video-MME score (74.5% vs 71.8%)

Qwen3 VL 8B Thinking

View details

Alibaba Cloud / Qwen Team

Larger context window (262,144 tokens)

Less expensive input tokens

Higher AIME 2025 score (80.3% vs 69.3%)

Higher BLINK score (68.7% vs 67.7%)

Higher CharXiv-D score (85.9% vs 85.5%)

Higher CharXiv-R score (53.0% vs 48.9%)

Higher DocVQAtest score (95.3% vs 95.0%)

Higher ERQA score (46.8% vs 43.0%)

Higher Hallusion Bench score (65.4% vs 61.5%)

Higher HMMT25 score (60.6% vs 50.6%)

Higher InfoVQAtest score (86.0% vs 82.0%)

Higher LiveBench 20241125 score (69.8% vs 65.4%)

Higher LiveCodeBench v6 score (58.6% vs 42.6%)

Higher MathVision score (62.7% vs 60.2%)

Higher MathVista-Mini score (81.4% vs 80.1%)

Higher MMBench-V1.1 score (87.5% vs 87.0%)

Higher MMLU score (85.2% vs 85.0%)

Higher MMLU-Redux score (88.8% vs 88.4%)

Higher MMStar score (75.3% vs 72.1%)

Higher MuirBench score (76.8% vs 62.9%)

Higher Multi-IF score (75.1% vs 66.1%)

Higher OCRBench-V2 (en) score (63.9% vs 63.2%)

Higher OCRBench-V2 (zh) score (59.2% vs 57.8%)

Higher OSWorld score (33.9% vs 30.3%)

Higher PolyMATH score (47.5% vs 44.3%)

Higher SimpleQA score (49.6% vs 27.0%)

Higher VideoMMMU score (72.8% vs 68.7%)

Higher WritingBench score (85.5% vs 82.6%)

Detailed Comparison

AI Model Comparison Table
Feature	Qwen3 VL 30B A3B Instruct	Qwen3 VL 8B Thinking