Is GPT-5 mini or Gemini 2.5 Flash cheaper?

GPT-5 mini is cheaper on all three of the workload shapes tested. On a RAG workload of 12k input and 700 output tokens at 50,000 requests per month with 60% caching, GPT-5 mini costs $139.00 per month and Gemini 2.5 Flash costs $170.30. Headline rates are $0.25/$2 versus $0.3/$2.5 per million input/output tokens.

GPT-5 mini vs Gemini 2.5 Flash

GPT-5 mini is the cheaper of the two across every workload shape tested below. Headline rates are $0.25 in / $2 out for GPT-5 mini versus $0.3 in / $2.5 out for Gemini 2.5 Flash, per million tokens.

Real monthly cost, three workloads

Workload	GPT-5 mini	Gemini 2.5 Flash	Cheaper
Support chatbot 1.5k in / 400 out · 100k req/mo · 30% cached	$107.38	$132.85	GPT-5 mini (1.2× cheaper)
RAG / doc Q&A 12k in / 700 out · 50k req/mo · 60% cached	$139.00	$170.30	GPT-5 mini (1.2× cheaper)
Coding agent 25k in / 2.5k out · 20k req/mo · 80% cached	$135.00	$167.00	GPT-5 mini (1.2× cheaper)

Costs include prompt caching at the stated hit rate and per-model tokenizer correction. Adjust for your own traffic →

Specifications side by side

	GPT-5 mini	Gemini 2.5 Flash
Provider	OpenAI	Google
Input / M tokens	$0.25	$0.3
Output / M tokens	$2	$2.5
Cached input	$0.025	$0.03
Context window	—	1000K
Batch discount	−50%	−50%
Tokenizer factor	×1	×1
Min prefix to cache	—	—
Announced retirement	—	—

Which should you pick?

On cost alone, GPT-5 mini wins every workload shape above.

Full detail: GPT-5 mini · Gemini 2.5 Flash

All figures verified 2026-08-18 against official provider documentation. Fields a provider does not publish are shown as “—” rather than estimated.