GPT-4.1 nano on Run BiOS
gpt-4.1-nano
Top up $20 and we add 50%. That gives you about 100.0M input tokens extra on GPT-4.1 nano, on us.
Create an accountPricing
USD per 1M tokens| Input | $0.10 |
| Output | $0.40 |
| Cached input | $0.0375% below fresh input |
Billed per token with no minimum and no monthly fee. Price your workload
Capabilities and limits
- Context window
- 1M tokens
- Max output
- 32K tokens
- Modality
- text
- Tool calling
- Supported
- Vision
- Not supported
- Reasoning
- Not supported
- Prompt caching
- Supported (automatic)
Measured performance
7d- Time to first token (p50)
- 400 ms
- Time to first token (p95)
- 800 ms
- Throughput (avg)
- 128.3 tok/s
- Throughput (p50)
- 150.0 tok/s
Measured on real Run BiOS traffic over the trailing 7d. Your figures will vary with prompt shape and region.
curl https://api.runbios.ai/v1/chat/completions \
-H "Authorization: Bearer $BIOS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-4.1-nano",
"messages": [{ "role": "user", "content": "Hello" }]
}'Existing OpenAI SDK code works by changing the base URL and the key. API overview
Model card
Curated by Run BiOS · benchmark figures, where present, are the card’s own, not our measurements
GPT-4.1 nano is the fastest, most cost-efficient version of GPT-4.1. It excels at instruction following and tool calling, and features a 1M-token context window and low latency without a reasoning step. For more complex tasks, OpenAI recommends starting with GPT-5 nano.
Model details
| Property | Value |
|---|---|
| Developer | OpenAI |
| Model ID | gpt-4.1-nano |
| OpenAI default snapshot | gpt-4.1-nano-2025-04-14 |
| Input | Text |
| Output | Text |
| Context window | 1,047,576 tokens |
| Max output tokens | 32,768 |
| Knowledge cutoff | Jun 01, 2024 |
| Reasoning | No (instruction model, no reasoning step) |
Details above are from OpenAI's official model page; context window and input types are as served on Run BiOS.
On Run BiOS
- No reasoning step. A
reasoning_effortsetting is accepted and ignored, so requests written for reasoning models still work. - Tool calling. Function tools with
tool_choiceauto, required, a named function or none, including parallel tool calls. - Structured outputs.
response_formatwithjson_schemaorjson_object. - Text input only on Run BiOS.
- Sampling parameters.
temperature,top_p,frequency_penalty,presence_penalty,stopandseedare supported. - Prompt caching. Automatic for repeated prompt prefixes; the minimum cacheable length varies with the request. Cache reads bill at the cached-input rate, and there is no separate charge for cache writes. Send a stable
prompt_cache_keyto improve cache hits. - Two APIs. Works with the OpenAI Chat Completions API (
/v1/chat/completions) and the Anthropic Messages API (/v1/messages).