Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows
发布日期2026年7月9日
参数规模—
上下文长度1.0M
许可证Proprietary
知识截止2026年2月16日
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| GPQA Diamond | general | 94.6 | 来源 |
| BrowseComp | reasoning search | 90.4 | 来源 |
| FrontierMath | math reasoning | 89.0 | 来源 |
| Terminal-Bench | reasoning code | 88.8 | 来源 |
| MMMU Pro | vision multimodal reasoning general | 83.0 | 来源 |
| Artificial Analysis Coding Agent Index | code | 80.0 | 来源 |
| DeepSWE | code | 72.7 | 来源 |
| SWE-Bench Pro | code | 64.6 | 来源 |
| OSWorld | multimodal general vision | 62.6 | 来源 |
| Artificial Analysis Intelligence Index | general | 58.9 | 来源 |
| Toolathlon | agent | 58.0 | 来源 |
| Agents' Last Exam | agent | 52.7 | 来源 |
Pricing
API 价格对比
| 服务商 | 输入价 | 输出价 | 上下文 | 吞吐(tok/s) | 延迟(s) | 函数调用 | 代码执行 | 联网搜索 |
|---|---|---|---|---|---|---|---|---|
| routing.run | $2.50 | $15.00 | 922K | — | — | ✓ | ✗ | ✗ |
| Abacus | $5.00 | $30.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| AI-ROUTER | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| AIHubMix | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| Azure | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| Azure Cognitive Services | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| GitHub Copilot | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| LLM Gateway | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| OpenAI | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| OpenCode Zen | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
| Vivgrid | $5.00 | $30.00 | 922K | — | — | ✓ | ✗ | ✗ |
价格单位：美元/百万 token，数据来自社区整理，仅供参考。