← Back to blog

Cheapest GPT-5 API: Complete Cost Breakdown (May 2026)

OpenAI's GPT-5 family now spans six models from $0.08 to $30 per million tokens. If you're looking for the cheapest GPT-5 API, you have more options than ever — but picking the wrong one can cost you 375x more than necessary.

We break down every GPT-5 family model by price, quality, and use case so you can find the cheapest option that actually works for your workload.

Every GPT-5 Family Model, Ranked by Price

Model	Input ($/1M tokens)	Output ($/1M tokens)	Context	Tier
GPT-oss 20B	$0.08	$0.35	128K	Budget
GPT-oss 120B	$0.15	$0.60	128K	Budget
GPT-4o mini	$0.15	$0.60	128K	Budget
GPT-5 mini	$0.25	$2.00	272K	Budget
GPT-5	$1.25	$10.00	272K	Premium
GPT-5.3 Codex	$1.75	$14.00	400K	Mid
GPT-5.5	$5.00	$30.00	1M	Premium
GPT-5.5 Pro	$30.00	$180.00	1M	Premium

The cheapest GPT-5 model (GPT-oss 20B at $0.08) costs 375x less than the most expensive (GPT-5.5 Pro at $30). That's the difference between $7/month and $2,625/month for a 10M token workload.

The Four Budget Tiers

Ultra-Budget: GPT-oss 20B ($0.08/$0.35)

OpenAI's open-weight model is the absolute cheapest GPT-5 option. At $0.08 per million input tokens, it costs less than a penny per request. Best for high-volume, simple tasks: classification, extraction, sentiment analysis, chatbot routing.

Budget: GPT-oss 120B / GPT-4o mini ($0.15/$0.60)

Both priced identically, these are workhorses for everyday tasks. GPT-oss 120B runs on open weights (self-host or via providers like Together.ai), while GPT-4o mini is a hosted API with built-in safety. Good for summarization, Q&A, content generation.

Mid-Range: GPT-5 mini ($0.25/$2.00)

The sweet spot for most developers. GPT-5 mini offers near-GPT-5 quality at a fraction of the cost. With 272K context, it handles complex multi-turn conversations and code generation. 67% cheaper than GPT-5.

Premium: GPT-5 ($1.25/$10.00)

Full GPT-5 for when you need maximum reasoning. Reserve this for the 10-20% of requests that genuinely require it — not every API call.

Monthly Cost Scenarios

At scale, GPT-5 mini offers 80% of GPT-5's quality at 80% lower cost. For most production workloads, that's the right tradeoff.

Competitor Alternatives Even Cheaper Than GPT-oss

If absolute lowest cost is your priority, models outside the GPT-5 family can beat even GPT-oss pricing:

Model	Input ($/1M)	Output ($/1M)	vs GPT-oss 20B
Gemini 2.0 Flash Lite	$0.075	$0.30	6% cheaper input
Llama 3.1 8B	$0.10	$0.10	25% cheaper output
DeepSeek V4 Flash	$0.14	$0.28	20% cheaper output
Gemini 2.0 Flash	$0.10	$0.40	13% cheaper input

Gemini 2.0 Flash Lite is the only API cheaper than GPT-oss 20B on input tokens — but GPT-oss models often produce better output quality for the same price.