Compare models
Pick up to 4 models for a side-by-side spec sheet. Hover any dotted metric name for what the test measures. Best value in each row is highlighted.
Gemini 3.6 Flashgpt-oss-120b
| Gemini 3.6 Flash | gpt-oss-120b OpenAI | |
|---|---|---|
| Pricing | ||
| Input $/1M | $1.50 | $0.037 |
| Output $/1M | $7.50 | $0.17 |
| Blended $/1M (3:1) | $3.00 | $0.07 |
| Cache read $/1M | $0.15 | — |
| Cache write $/1M | $0.083 | — |
| Quality indices | ||
| Intelligence Index | 50.1 | 23.8 |
| Coding Index | 69.2 | 30.4 |
| Math Index | — | 93.4 |
| Agentic Index | 38.7 | 13.2 |
| Speed & latency | ||
| Output speed (tok/s) | 311 | 292 |
| Time to first token | 12.80s | 0.50s |
| Time to first answer token | 12.80s | 7.35s |
| Benchmarks (Artificial Analysis) | ||
| MMLU-Pro | — | 80.8% |
| GPQA Diamond | 92.8% | 78.2% |
| Humanity's Last Exam | 38.3% | 18.5% |
| LiveCodeBench | — | 87.8% |
| SciCode | 52.7% | 38.9% |
| AIME 2025 | — | 93.4% |
| IFBench | — | 69.0% |
| AA-LCR (Long Context Reasoning) | 69.7% | 50.7% |
| Terminal-Bench Hard | — | 23.5% |
| Terminal-Bench 2.1 | 77.5% | 26.2% |
| τ²-Bench (Telecom) | — | 65.8% |
| τ³-Bench Banking | 24.5% | 12.0% |
| Design Arena (Elo) | ||
| Website | 1,332 | 995 |
| UI components | — | 966 |
| Data viz | — | 1,037 |
| 3D | — | 959 |
| Game dev | — | 1,052 |
| Code categories | — | 997 |
| Specs | ||
| Context window | 1.05M | 131K |
| Max output tokens | 66K | 131K |
| Input modalities | text, image, video, file, audio | text |
| Tokenizer | Gemini | GPT |