GPT-5 Mini on Run BiOS

gpt-5-mini

+50%First top-up bonus

Top up $20 and we add 50%. That gives you about 40.0M input tokens extra on GPT-5 Mini, on us.

Create an account

266K contextVisionTool callingReasoningPrompt caching

Pricing

USD per 1M tokens

Input	$0.25
Output	$2.00
Cached input	$0.0390% below fresh input

Billed per token with no minimum and no monthly fee. Price your workload

Capabilities and limits

Context window: 266K tokens
Max output: 125K tokens
Modality: multimodal
Tool calling: Supported
Vision: Supported
Reasoning: Always on
Reasoning effort: minimal, low, medium, high (default: medium)
Prompt caching: Supported (automatic)

Measured performance

Time to first token (p50): 3.0 s
Time to first token (p95): 3.0 s
Throughput (avg): 119.7 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.

Call it now: OpenAI-compatible

curl https://api.runbios.ai/v1/chat/completions \
  -H "Authorization: Bearer $BIOS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5-mini",
    "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-5 Mini is a faster, more cost-efficient version of GPT-5. It is great for well-defined tasks and precise prompts. For most new low-latency, high-volume workloads, OpenAI recommends starting with GPT-5.6 Terra.

Model details

Property	Value
Developer	OpenAI
Model ID	`gpt-5-mini`
OpenAI default snapshot	`gpt-5-mini-2025-08-07`
Input	Text, image
Output	Text
Context window	272,000 tokens on Run BiOS (OpenAI's full window is 400,000 tokens)
Max output tokens	128,000
Knowledge cutoff	May 31, 2024
Reasoning	Yes

Details above are from OpenAI's official model page; context window and input types are as served on Run BiOS.

On Run BiOS

Always reasons. Reasoning levels minimal, low, medium, high; the default is medium. This model cannot turn reasoning off: a request for no reasoning runs at the lowest available level.
Reasoning summaries. When reasoning is on, OpenAI may return a summary of the model's reasoning: in reasoning_content on the OpenAI API, and in thinking blocks on the Anthropic Messages API when thinking.display is summarized. OpenAI does not return a summary for every request.
Tool calling. Function tools with tool_choice auto, required, a named function or none, including parallel tool calls, and together with reasoning.
Structured outputs. response_format with json_schema or json_object.
Image input. Text and image input, text output.
Sampling parameters. temperature, top_p, penalties, stop and seed are not applied to this model on Run BiOS, because it always reasons; the request is still served.
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_key to improve cache hits.
Two APIs. Works with the OpenAI Chat Completions API (/v1/chat/completions) and the Anthropic Messages API (/v1/messages).