gpt-oss-20B (high) vs gpt-oss-20B (low)
Comparing 2 AI models · 5 benchmarks · OpenAI
Composite Indices
Intelligence, Coding, Math
Standard Benchmarks
Academic and industry benchmarks
Benchmark Winners
gpt-oss-20B (high)
- GPQA
- MMLU Pro
- HLE
- LiveCodeBench
- AIME 2025
gpt-oss-20B (low)
No clear wins
| Metric | Op gpt-oss-20B (high) | Op gpt-oss-20B (low) |
|---|---|---|
| Pricing Per 1M tokens | ||
| Input Cost | $0.07/1M | $0.07/1M |
| Output Cost | $0.20/1M | $0.20/1M |
| Blended Cost 3:1 input/output ratio | $0.10/1M | $0.10/1M |
| Specifications | ||
| Organization Model creator | OpenAI | OpenAI |
| Release Date Launch date | Aug 5, 2025 | Aug 5, 2025 |
| Performance & Speed | ||
| Throughput Output speed | 236.2 tok/s | 235.0 tok/s |
| Time to First Token (TTFT) Initial response delay | 586ms | 615ms |
| Latency Time to first answer token | 9054ms | 9125ms |
| Composite Indices | ||
| Intelligence Index Overall reasoning capability | 52.4 | 44.3 |
| Coding Index Programming ability | 40.7 | 34.5 |
| Math Index Mathematical reasoning | 89.3 | 62.3 |
| Standard Benchmarks | ||
| GPQA Graduate-level reasoning | 68.8% | 61.1% |
| MMLU Pro Advanced knowledge | 74.8% | 71.8% |
| HLE Hard language evaluation | 9.8% | 5.1% |
| LiveCodeBench Real-world coding tasks | 77.7% | 65.2% |
| MATH 500 Mathematical problems | — | — |
| AIME 2025 Advanced math competition | 89.3% | 62.3% |
| AIME (Original) Math olympiad problems | — | — |
| SciCode Scientific code generation | 34.4% | 34.0% |
| LCR Code review capability | 34.3% | 31.0% |
| IFBench Instruction-following | 65.1% | 57.8% |
| TAU-bench v2 Tool use & agentic tasks | 60.2% | 50.3% |
| TerminalBench Hard CLI command generation | 9.9% | 4.3% |
Key Takeaways
gpt-oss-20B (high) offers the best value at $0.07/1M, making it ideal for high-volume applications and cost-conscious projects.
gpt-oss-20B (high) leads in reasoning capabilities with a 68.8% GPQA score, excelling at complex analytical tasks and problem-solving.
gpt-oss-20B (high) achieves a 40.7 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-oss-20B (high)
- Cost-sensitive applications
- High-volume processing
- Complex reasoning tasks
- Research & analysis
- Code generation
- Software development
gpt-oss-20B (low)
- General-purpose AI
- Versatile applications
Try Models for Free
Try gpt-oss-20B (high) for FREE
No credit card or account required.
Try gpt-oss-20B (low) for FREE
No credit card or account required.
Cost Calculator
gpt-oss-20B (high) gpt-oss-20B (low) Costs are estimates based on API pricing. Actual costs may vary based on caching, batch processing, and volume discounts.
Recommended Comparisons
