Anthropic: Claude Sonnet 5
Chatanthropic/claude-sonnet-5Claude Sonnet 5 is Anthropic's flagship Sonnet-series model released on 2026-07-01, featuring a 1M-token context window and 128K output. It supports adaptive thinking (extended reasoning), vision, tool use, prompt caching, and web search. Available via both OpenAI-compatible and native Anthropic protocols through Ofox.
Providers
provider.type: "anthropic"Documentation
How to call Claude Sonnet 5 on OfoxAI: model IDs, request parameters and the errors you are most likely to hit.
- Claude API guideModel IDs, choosing an endpoint, code examples and common errors.
- Thinking and reasoning settingsTurn thinking on or off, set the effort or token budget, and read the reasoning output.
- Prompt cachingMark cache breakpoints, choose a cache lifetime and confirm cache hits.
- Long context and PDF inputSend long documents and PDF files, and what counts against the context window.
Code Examples
import anthropicclient = anthropic.Anthropic(base_url="https://api.ofox.ai/anthropic",api_key="YOUR_OFOX_API_KEY",)message = client.messages.create(model="anthropic/claude-sonnet-5",max_tokens=1024,messages=[{"role": "user", "content": "Hello!"}],)print(message.content[0].text)
Uptime & Status
Benchmarks
Anthropic: Claude Sonnet 5 scores 1463 in the Overall category of the LMArena text leaderboard (style control), ranking #36 of 374 models based on 9,850 human preference votes (updated 2026-07-12).
| Category | Arena Score | 95% CI | Rank | Votes |
|---|---|---|---|---|
| Overall | 1463 | 1456–1470 | #36 of 374 | 9,850 |
| Hard Prompts | 1493 | 1485–1501 | #23 of 374 | 6,491 |
| Coding | 1522 | 1510–1534 | #15 of 369 | 2,725 |
| Math | 1463 | 1436–1490 | #36 of 362 | 449 |
| Creative Writing | 1430 | 1416–1445 | #48 of 372 | 1,699 |
| Instruction Following | 1465 | 1454–1476 | #22 of 374 | 3,269 |
| Chinese | 1509 | 1482–1535 | #23 of 344 | 499 |
Source: LMArena · CC BY 4.0 · Updated 2026-07-12 · Methodology ↗ · Ranks compare models within each category of the LMArena text leaderboard (style control). Scores come from third-party human preference evaluations, not from OFOX.