Model Comparison
Qwen3 8B (Non-reasoning)
vs. Qwen3 VL 235B A22B (Reasoning)
Comparing 2 AI models · 12 benchmarks · Alibaba
Recommended Pick
Al Qwen3 VL 235B A22B (Reasoning) 14 metric wins Strongest on: Reasoning, Intelligence, Coding
Best Value
Qwen3 8B (Non-reasoning)
100.0 value score
30.6 reasoning / $0.18/1M
Lowest Price
Qwen3 8B (Non-reasoning)
$0.18/1M input price
Best Reasoning
Qwen3 VL 235B A22B (Reasoning)
58.3 reasoning score
Blends available reasoning benchmarks
Best for Coding
Qwen3 VL 235B A22B (Reasoning)
20.9 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Best value
Qwen3 8B (Non-reasoning) has the strongest quality-to-price mix at 100.0 out of 100 value points.
Price gap
Qwen3 8B (Non-reasoning) is 4.7x cheaper on input tokens than Qwen3 VL 235B A22B (Reasoning).
Speed gap
Qwen3 8B (Non-reasoning) generates about 1.3x as many tokens per second as Qwen3 VL 235B A22B (Reasoning).
Reasoning gap
Qwen3 VL 235B A22B (Reasoning) leads Qwen3 8B (Non-reasoning) by 27.7 points on reasoning.
Coding gap
Qwen3 VL 235B A22B (Reasoning) leads Qwen3 8B (Non-reasoning) by 13.8 points on coding.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
Qwen3 8B (Non-reasoning)
Alibaba
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Qwen3 VL 235B A22B (Reasoning)
Alibaba
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | Al Qwen3 8B (Non-reasoning) | Top Pick Al Qwen3 VL 235B A22B (Reasoning) |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.18/1M | $0.84/1M |
| Output Cost | $0.20/1M | $6.17/1M |
| Blended (3:1) | $0.18/1M | $2.17/1M |
| Specifications | ||
| Organization | Alibaba | Alibaba |
| Release Date | Apr 28, 2025 | Sep 23, 2025 |
| Performance & Speed | ||
| Throughput | 64.0 tok/s | 48.6 tok/s |
| TTFT | 1293ms | 3810ms |
| Latency | 1293ms | 44958ms |
| Composite Indices | ||
| Value Score | 100.0 | 16.2 |
| Reasoning Score | 30.6 | 58.3 |
| Intelligence | 10.6 | 27.6 |
| Coding | 7.1 | 20.9 |
| Math | 24.3 | 88.3 |
| Standard Benchmarks | ||
| GPQA | 45.2% | 77.2% |
| MMLU Pro | 64.3% | 83.6% |
| HLE | 2.8% | 10.1% |
| LiveCodeBench | 20.2% | 64.6% |
| MATH 500 | 82.8% | — |
| AIME 2025 | 24.3% | 88.3% |
| AIME (Original) | 24.3% | — |
| SciCode | 16.8% | 39.9% |
| LCR | 0.0% | 58.7% |
| IFBench | 28.6% | 56.5% |
| TAU-bench v2 | 24.9% | 54.1% |
| TerminalBench Hard | 2.3% | 11.4% |
Key Takeaways
Qwen3 8B (Non-reasoning) offers the best value at $0.18/1M, making it ideal for high-volume applications and cost-conscious projects.
Qwen3 VL 235B A22B (Reasoning) has the strongest reasoning profile with a 58.3 reasoning score, combining the available reasoning-heavy benchmarks.
Qwen3 VL 235B A22B (Reasoning) reaches a 20.9 coding index, making it the top choice for software development and code generation tasks.
All models support context windows of ∞+ tokens, suitable for processing lengthy documents and maintaining extended conversations.
When to Choose Each Model
Qwen3 8B (Non-reasoning)
- Cost-sensitive applications
- High-volume processing
Qwen3 VL 235B A22B (Reasoning)
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
