GPT-5 mini (medium) vs GPT-5 mini (high)
Comparing 2 AI models · 5 benchmarks · OpenAI
Composite Indices
Intelligence, Coding, Math
Standard Benchmarks
Academic and industry benchmarks
Benchmark Winners
GPT-5 mini (medium)
No clear wins
GPT-5 mini (high)
- GPQA
- MMLU Pro
- HLE
- LiveCodeBench
- AIME 2025
| Metric | Op GPT-5 mini (medium) | Op GPT-5 mini (high) |
|---|---|---|
| Pricing Per 1M tokens | ||
| Input Cost | $0.25/1M | $0.25/1M |
| Output Cost | $2.00/1M | $2.00/1M |
| Blended Cost 3:1 input/output ratio | $0.69/1M | $0.69/1M |
| Specifications | ||
| Organization Model creator | OpenAI | OpenAI |
| Release Date Launch date | Aug 7, 2025 | Aug 7, 2025 |
| Performance & Speed | ||
| Throughput Output speed | 73.2 tok/s | — |
| Time to First Token (TTFT) Initial response delay | 28690ms | — |
| Latency Time to first answer token | 28690ms | — |
| Composite Indices | ||
| Intelligence Index Overall reasoning capability | 60.8 | 64.3 |
| Coding Index Programming ability | 45.7 | 51.4 |
| Math Index Mathematical reasoning | 85.0 | 90.7 |
| Standard Benchmarks | ||
| GPQA Graduate-level reasoning | 80.3% | 82.8% |
| MMLU Pro Advanced knowledge | 82.8% | 83.7% |
| HLE Hard language evaluation | 14.6% | 19.7% |
| LiveCodeBench Real-world coding tasks | 69.2% | 83.8% |
| MATH 500 Mathematical problems | — | — |
| AIME 2025 Advanced math competition | 85.0% | 90.7% |
| AIME (Original) Math olympiad problems | — | — |
| SciCode Scientific code generation | 41.0% | 39.2% |
| LCR Code review capability | 66.0% | 68.0% |
| IFBench Instruction-following | 71.2% | 75.4% |
| TAU-bench v2 Tool use & agentic tasks | 71.1% | 68.4% |
| TerminalBench Hard CLI command generation | 27.0% | 31.2% |
Key Takeaways
GPT-5 mini (medium) offers the best value at $0.25/1M, making it ideal for high-volume applications and cost-conscious projects.
GPT-5 mini (high) leads in reasoning capabilities with a 82.8% GPQA score, excelling at complex analytical tasks and problem-solving.
GPT-5 mini (high) achieves a 51.4 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
GPT-5 mini (medium)
- Cost-sensitive applications
- High-volume processing
GPT-5 mini (high)
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
Try Models for Free
Try GPT-5 mini (medium) for FREE
No credit card or account required.
Try GPT-5 mini (high) for FREE
No credit card or account required.
Cost Calculator
GPT-5 mini (medium) GPT-5 mini (high) Costs are estimates based on API pricing. Actual costs may vary based on caching, batch processing, and volume discounts.