Claude Opus 4.8
AnthropicAnthropic's top model for complex agentic and long-horizon tasks.
✓ Best for
Long-horizon agents, complex multi-step tasks, careful writing
✗ Not ideal for
High-throughput cheap calls
1M+ contextFrontier intelligenceMultimodal vision
TextVisionTool useCode
Performance
Artificial Analysis — Intelligence Index· 2026-06-17| Intelligence Index | 61Frontier · /100 |
| Coding | 90 |
| Math | 88 |
| Reasoning | 91 |
| Vision | 82 |
| Arena Elo | 1475Top-tier |
Specifications
| Vendor | Anthropic |
| Family | Claude |
| Context | 1M · 750k words（≈ a novella） |
| Max output | 128K · 96k words（≈ a thesis）Below average |
| Released | 2026-04-16 |
| Open source | No |
| Tool use | ✓ |
Technical specs (developers: rate limits / latency / throughput)
| Rate limit (RPM) | 100 |
| Rate limit (TPM) | 50,000 |
| TTFT | 520 ms |
| Output speed | 65 tok/s |
Pricing across channels
per 1M tokens, USD
Anthropic Pricing· 2026-06-17
Related models
Similar tier, vendor or price range