Model Comparison
Qwen3 VL 30B A3B Instruct
vs. Tri-21B-think Preview
Comparing 2 AI models · 10 benchmarks · Alibaba, Trillion Labs
Recommended Pick
Al Qwen3 VL 30B A3B Instruct 8 metric wins Strongest on: Throughput, Reasoning, Coding
Lowest Price
Tri-21B-think Preview
$0.00/1M input price
Best Reasoning
Qwen3 VL 30B A3B Instruct
47.3 reasoning score
Blends available reasoning benchmarks
Best for Coding
Qwen3 VL 30B A3B Instruct
14.3 coding index
Composite Indices
Higher is better; speed and price are normalized
Standard Benchmarks
Only benchmarks with data are shown
Differences That Matter
Price gap
Tri-21B-think Preview is ∞x cheaper on input tokens than Qwen3 VL 30B A3B Instruct.
Reasoning gap
Qwen3 VL 30B A3B Instruct leads Tri-21B-think Preview by 20.8 points on reasoning.
Coding gap
Qwen3 VL 30B A3B Instruct leads Tri-21B-think Preview by 6.9 points on coding.
Top-pick rationale
Qwen3 VL 30B A3B Instruct wins 8 measurable categories, including Throughput, Reasoning, Coding, GPQA.
Response Face-Off
Run one prompt through the selected models and compare response quality with live speed and cost context.
Qwen3 VL 30B A3B Instruct
Alibaba
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Tri-21B-think Preview
Trillion Labs
TTFT
—
Time
—
tok/s
—
Tokens
—
Cost
—
Which answer was more useful?
Full Comparison
| Metric | Top Pick Al Qwen3 VL 30B A3B Instruct | Tr Tri-21B-think Preview |
|---|---|---|
| Pricing per 1M tokens | ||
| Input Cost | $0.20/1M | $0.00/1M |
| Output Cost | $0.60/1M | $0.00/1M |
| Blended (3:1) | $0.30/1M | — |
| Specifications | ||
| Organization | Alibaba | Trillion Labs |
| Release Date | Oct 3, 2025 | Feb 10, 2026 |
| Performance & Speed | ||
| Throughput | 120.9 tok/s | — |
| TTFT | 1089ms | — |
| Latency | 1089ms | — |
| Composite Indices | ||
| Value Score | 100.0 | — |
| Reasoning Score | 47.3 | 26.5 |
| Intelligence | 16.0 | 20.0 |
| Coding | 14.3 | 7.4 |
| Math | 72.3 | — |
| Standard Benchmarks | ||
| GPQA | 69.5% | 53.8% |
| MMLU Pro | 76.4% | — |
| HLE | 6.4% | 5.7% |
| LiveCodeBench | 47.6% | — |
| AIME 2025 | 72.3% | — |
| SciCode | 30.8% | 17.8% |
| LCR | 23.7% | 14.7% |
| IFBench | 33.1% | 47.1% |
| TAU-bench v2 | 19.0% | 93.3% |
| TerminalBench Hard | 6.1% | 2.3% |
Key Takeaways
Tri-21B-think Preview offers the best value at $0.00/1M, making it ideal for high-volume applications and cost-conscious projects.
Qwen3 VL 30B A3B Instruct has the strongest reasoning profile with a 47.3 reasoning score, combining the available reasoning-heavy benchmarks.
Qwen3 VL 30B A3B Instruct reaches a 14.3 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 VL 30B A3B Instruct
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
Tri-21B-think Preview
- Cost-sensitive applications
- High-volume processing
