Sim is the open-source AI workspace where teams build, deploy, and manage AI agents. This page compares GPT-5 and Claude Sonnet 4.5 on API pricing per 1M tokens, context window, max output, and capabilities. Both models are available in the Sim Agent block, and each Agent block in a Sim agent can use a different model.
API model comparison
GPT-5 vs Claude Sonnet 4.5
GPT-5 is 2× cheaper than Claude Sonnet 4.5: 1,000 agent calls with 10k input and 1k output tokens each cost $22.5 versus $45 at standard rates.
GPT-5 charges $1.25 input and $10 output per 1M tokens; Claude Sonnet 4.5 charges $3 and $15.
GPT-5 has a 2× larger context window: 400k versus 200k tokens.
GPT-5 can generate up to 128k output tokens per response, versus 64k.
Only Claude Sonnet 4.5 supports native structured outputs in Sim.
Pricing, context window, and capabilities
API prices are per 1M tokens. Pricing last updated .
| Metric | GPT-5 | Claude Sonnet 4.5 |
|---|---|---|
| Provider | OpenAI | Anthropic |
| Input price | $1.25/1M | $3/1M |
| Cached input | $0.125/1M | $0.3/1M |
| Output price | $10/1M | $15/1M |
| Context window | 400k tokens | 200k tokens |
| Released | Aug 7, 2025 | Sep 29, 2025 |
| Temperature | Not configurable | 0 to 1 |
| Reasoning effort | minimal, low, medium, high | Not supported |
| Verbosity | low, medium, high | Not supported |
| Thinking levels | Not supported | low, medium, high (default: high) |
| Structured outputs | Supported | Supported (native) |
| Tool choice | Supported | Supported |
| Computer use | Not supported | Not supported |
| Deep research | Not supported | Not supported |
| Memory support | Supported | Supported |
| Max output tokens | 128k | 64k |
Use GPT-5 or Claude Sonnet 4.5 in a Sim agent
Both models are available in the Sim Agent block. Pick either one from the model picker, and each Agent block in the same agent can use a different model.
Frequently asked questions
GPT-5 is 2× cheaper than Claude Sonnet 4.5: 1,000 agent calls with 10k input and 1k output tokens each cost $22.5 versus $45 at standard rates. GPT-5 charges $1.25 input and $10 output per 1M tokens; Claude Sonnet 4.5 charges $3 and $15.
