DeepSeek R1 0528 Qwen3 8B vs DeepSeek R1 Distill Qwen 14B
Comparing 2 AI models · 6 benchmarks · DeepSeek
Most Affordable
De
DeepSeek R1 0528 Qwen3 8B
$0.06/1M
Highest Intelligence
De
DeepSeek R1 0528 Qwen3 8B
61.2% GPQA
Best for Coding
De
DeepSeek R1 0528 Qwen3 8B
24.4 Coding Index
Price Difference
2.5x
input cost range
Composite Indices
Intelligence, Coding, Math
Standard Benchmarks
Academic and industry benchmarks
Benchmark Winners
6 tests
De
DeepSeek R1 0528 Qwen3 8B
4
- GPQA
- HLE
- LiveCodeBench
- AIME 2025
De
DeepSeek R1 Distill Qwen 14B
2
- MMLU Pro
- MATH 500
| Metric | De DeepSeek R1 0528 Qwen3 8B | De DeepSeek R1 Distill Qwen 14B |
|---|---|---|
| Pricing Per 1M tokens | ||
| Input Cost | $0.06/1M | $0.15/1M |
| Output Cost | $0.09/1M | $0.15/1M |
| Blended Cost 3:1 input/output ratio | $0.07/1M | $0.15/1M |
| Specifications | ||
| Organization Model creator | DeepSeek | DeepSeek |
| Release Date Launch date | May 29, 2025 | Jan 20, 2025 |
| Performance & Speed | ||
| Throughput Output speed | 68.3 tok/s | 47.5 tok/s |
| Time to First Token (TTFT) Initial response delay | 864ms | 1132ms |
| Latency Time to first answer token | 30132ms | 43258ms |
| Composite Indices | ||
| Intelligence Index Overall reasoning capability | 31.0 | 29.7 |
| Coding Index Programming ability | 24.4 | — |
| Math Index Mathematical reasoning | 63.7 | 55.7 |
| Standard Benchmarks | ||
| GPQA Graduate-level reasoning | 61.2% | 48.4% |
| MMLU Pro Advanced knowledge | 73.9% | 74.0% |
| HLE Hard language evaluation | 5.6% | 4.4% |
| LiveCodeBench Real-world coding tasks | 51.3% | 37.6% |
| MATH 500 Mathematical problems | 93.2% | 94.9% |
| AIME 2025 Advanced math competition | 63.7% | 55.7% |
| AIME (Original) Math olympiad problems | 65.0% | 66.7% |
| SciCode Scientific code generation | 20.4% | 23.9% |
| LCR Code review capability | 13.0% | 7.0% |
| IFBench Instruction-following | 19.9% | 22.1% |
| TAU-bench v2 Tool use & agentic tasks | 0.0% | — |
| TerminalBench Hard CLI command generation | 1.4% | — |
Key Takeaways
DeepSeek R1 0528 Qwen3 8B offers the best value at $0.06/1M, making it ideal for high-volume applications and cost-conscious projects.
DeepSeek R1 0528 Qwen3 8B leads in reasoning capabilities with a 61.2% GPQA score, excelling at complex analytical tasks and problem-solving.
DeepSeek R1 0528 Qwen3 8B achieves a 24.4 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
De
DeepSeek R1 0528 Qwen3 8B
- Cost-sensitive applications
- High-volume processing
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
De
DeepSeek R1 Distill Qwen 14B
- General-purpose AI
- Versatile applications
Try Models for Free
De