GPT-5 API Pricing Comparison 2026: GPT-5.4 vs GPT-5.2 vs GPT-5.1
GPT-5 API Pricing Comparison: GPT-5.4 vs GPT-5.2 vs GPT-5.1
This guide covers exact per-token pricing for every GPT-5.x model, cached input rates, EvoLink discount pricing, and how to choose the right model for your budget.
GPT-5 API Pricing Comparison
| Model | Input (per 1M) | Output (per 1M) | Cached Input (per 1M) | Context | Max Output |
|---|---|---|---|---|---|
| GPT-5.4 | $2.50 | $15.00 | $0.25 | 1.05M | 128K |
| GPT-5.2 | $1.75 | $14.00 | $0.175 | 400K | 128K |
| GPT-5.1 | $1.25 | $10.00 | $0.125 | 400K | 128K |
All prices are per 1M tokens at OpenAI base rates. EvoLink offers discounted pricing on GPT-5.4 (see below).
GPT-5.4 Pricing Details
GPT-5.4 is OpenAI's latest flagship model with 1.05M context, native computer use, and Tool Search.
| Tier | Input | Output | Cached Input |
|---|---|---|---|
| Base rate | $2.50 / 1M | $15.00 / 1M | $0.25 / 1M |
| EvoLink (20% off) | $2.00 / 1M | $12.00 / 1M | $0.20 / 1M |
| >272K input tier | $5.00 / 1M | $22.50 / 1M | — |
GPT-5.2 Pricing Details
GPT-5.2 is the production workhorse — strong reasoning and coding at a lower input price.
| Tier | Input | Output | Cached Input |
|---|---|---|---|
| Base rate | $1.75 / 1M | $14.00 / 1M | $0.175 / 1M |
GPT-5.1 Pricing Details
GPT-5.1 is the budget option for simpler tasks that don't need GPT-5.2's reasoning depth.
| Tier | Input | Output | Cached Input |
|---|---|---|---|
| Base rate | $1.25 / 1M | $10.00 / 1M | $0.125 / 1M |
Cached Input Pricing: 90% Savings Across All GPT-5 Models
All GPT-5.x models offer a 90% discount on cached input tokens. This matters because repeated system prompts and instructions often make up the majority of input cost.
| Model | Standard Input | Cached Input | Savings |
|---|---|---|---|
| GPT-5.4 | $2.50 / 1M | $0.25 / 1M | 90% |
| GPT-5.2 | $1.75 / 1M | $0.175 / 1M | 90% |
| GPT-5.1 | $1.25 / 1M | $0.125 / 1M | 90% |
GPT-5 vs Competitors: Price Comparison
| Model | Input / Output (per 1M) | Context | Best For |
|---|---|---|---|
| DeepSeek Chat | $0.27 / $1.10 | 64K | Budget tasks, high volume |
| Gemini 2.5 Flash | $0.30 / $2.50 | 1M | Fast, cheap, long-context |
| GPT-5.1 | $1.25 / $10.00 | 400K | Value-tier flagship |
| Gemini 3.1 Pro | $2.00 / $12.00 | 1M | Multimodal, long-context |
| GPT-5.2 | $1.75 / $14.00 | 400K | Production coding & reasoning |
| GPT-5.4 | $2.50 / $15.00 | 1.05M | Latest flagship, computer use |
| Claude Sonnet 4.6 | $3.00 / $15.00 | 1M | Coding and agentic tasks |
| Claude Opus 4.6 | $5.00 / $25.00 | 1M | Research and complex reasoning |
GPT-5.2 offers more output tokens per dollar than Claude Sonnet 4.6 ($14.00 vs $15.00) with comparable coding quality. GPT-5.4 through EvoLink ($12.00 output) beats both.
EvoLink GPT-5 Pricing: How It Works
EvoLink is a unified API gateway that gives you one API key for GPT-5.4, GPT-5.2, GPT-5.1, Claude, Gemini, and 200+ other models.
- GPT-5.4: 20% discount vs OpenAI direct ($2.00/$12.00 vs $2.50/$15.00)
- GPT-5.2 and GPT-5.1: Same pricing as OpenAI direct
- No subscription, no monthly minimum — pay per token
- 100% OpenAI SDK compatible — just change the base URL
Cost Optimization Tips
1. Route by task complexity
Not every request needs GPT-5.4. Send simple tasks to GPT-5.1 ($1.25 input) and reserve GPT-5.4 for complex reasoning, long-context analysis, and computer-use workflows.
2. Maximize cached input
With system prompts that stay constant, cached input pricing reduces your effective input cost by 10x. This is especially impactful on GPT-5.4 where standard input is $2.50/1M but cached is $0.25/1M.
3. Optimize for cost per task, not cost per token
A model that solves in one pass at $0.02 per call is cheaper than a model that needs 3 retries at $0.01 each. Track total cost per successful outcome.
4. Use EvoLink for GPT-5.4
EvoLink's 20% discount on GPT-5.4 means $12.00/1M output instead of $15.00. Over 1B output tokens, that saves $3,000.
FAQ
How much does GPT-5 API cost?
It depends on the version. GPT-5.4 costs $2.50 input / $15.00 output per 1M tokens. GPT-5.2 costs $1.75 / $14.00. GPT-5.1 costs $1.25 / $10.00. All offer 90% cached input discounts.
Which GPT-5 model is cheapest?
GPT-5.1 at $1.25/$10.00 per 1M tokens is the cheapest GPT-5 model. For the best price-performance ratio, GPT-5.2 at $1.75/$14.00 offers significantly better reasoning at only 40% more input cost.
Is GPT-5.4 worth the higher price?
GPT-5.4 is worth it if you need 1M+ context, computer-use capabilities, or the best available reasoning. Through EvoLink, GPT-5.4 output ($12.00/1M) is actually cheaper than GPT-5.2 base-rate output ($14.00/1M), making the upgrade more attractive.
How much does a typical GPT-5 API call cost?
A typical call with 2,000 input tokens and 500 output tokens costs: GPT-5.4 = ~$0.0125, GPT-5.2 = ~$0.0105, GPT-5.1 = ~$0.0075. With cached input, costs drop further.
Does EvoLink offer discounts on GPT-5?
EvoLink offers 20% off GPT-5.4 pricing ($2.00/$12.00 instead of $2.50/$15.00). GPT-5.2 and GPT-5.1 are at the same rate as OpenAI direct. No monthly minimum or subscription required.
