Sim is the open-source AI workspace where teams build, deploy, and manage AI agents. This page compares GPT-6 Luna and Gemini 3.8 Flash 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-6 Luna vs Gemini 3.8 Flash
GPT-6 Luna is 7.5× cheaper than Gemini 3.8 Flash: 1,000 agent calls with 10k input and 1k output tokens each cost $1.5 versus $11.25 at standard rates.
GPT-6 Luna charges $0.1 input and $0.5 output per 1M tokens; Gemini 3.8 Flash charges $0.75 and $3.75.
Their context windows are about the same size: 1.05M versus 1.05M tokens.
GPT-6 Luna can generate up to 128k output tokens per response, versus 66k.
Both support configurable reasoning and tool choice control in Sim.
Pricing, context window, and capabilities
API prices are per 1M tokens. Pricing last updated .
| Metric | GPT-6 Luna | Gemini 3.8 Flash |
|---|---|---|
| Provider | OpenAI | |
| Input price | $0.1/1M | $0.75/1M |
| Cached input | $0.01/1M | $0.075/1M |
| Output price | $0.5/1M | $3.75/1M |
| Long-context pricing | Above 272k input tokens, the full request is priced at $0.2/1M input tokens, $0.02/1M cached input tokens, and $0.75/1M output tokens. | No long-context tier |
| Context window | 1.05M tokens | 1.05M tokens |
| Released | Sep 22, 2026 | Sep 2, 2026 |
| Temperature | Not configurable | 0 to 2 |
| Reasoning effort | none, low, medium, high, xhigh, max | Not supported |
| Verbosity | low, medium, high | Not supported |
| Thinking levels | Not supported | low, medium, high (default: medium) |
| Structured outputs | Supported | Supported |
| Tool choice | Supported | Supported |
| Computer use | Not supported | Not supported |
| Deep research | Not supported | Not supported |
| Memory support | Supported | Supported |
| Max output tokens | 128k | 66k |
Use GPT-6 Luna or Gemini 3.8 Flash 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-6 Luna is 7.5× cheaper than Gemini 3.8 Flash: 1,000 agent calls with 10k input and 1k output tokens each cost $1.5 versus $11.25 at standard rates. GPT-6 Luna charges $0.1 input and $0.5 output per 1M tokens; Gemini 3.8 Flash charges $0.75 and $3.75.
