Use this when you need to choose the right OpenAI model tier for a production workflow using cost, quality, and latency—not intuition.
The fast answer: GPT-5.5 is the flagship; GPT-5-mini is the low-cost tier. The gap is 20x on input and 15x on output. For classification, extraction, and simple summarization, GPT-5-mini usually holds quality. For complex reasoning or high-stakes output, GPT-5.5 earns its cost. GPT-4o was retired in February 2026—this article focuses on the current GPT-5 family.
What you will get in 10 minutes
- Current OpenAI GPT-5 pricing (July 2026)
- When premium models are worth it vs when smaller models are enough
- A selection scorecard for one workflow
Use this when
- You are choosing a default model for a new feature
- Your OpenAI bill is growing and you want to test a cheaper tier
- You are migrating from GPT-4o or older models (retired)
- You need a cost vs quality framework, not vendor marketing
Current OpenAI lineup (July 2026)
OpenAI's production text models are now in the GPT-5 family. GPT-4o was retired in ChatGPT in February 2026; API access may remain for legacy integrations but is not the default. GPT-5.5 (and GPT-5.5 Pro) are the current GA flagship line, with the GPT-5.6 line (Sol / Terra / Luna) in limited preview. Plan for GPT-5. Current rates for these models and every other major provider are in the LLM API Pricing Index, updated daily.
| Model | Input ($/1M) | Output ($/1M) | Cached input ($/1M) | Context | Best for |
|---|---|---|---|---|---|
| GPT-5.5 | $5.00 | $30.00 | — | 1M | Complex reasoning, high-stakes output |
| GPT-5.5 Pro | $30.00 | $180.00 | — | 1M | Highest-effort reasoning |
| GPT-5-mini | $0.25 | $2.00 | $0.025 | 400K | Classification, extraction, simple tasks |
Prompts over the standard context threshold can be billed at higher long-context rates. Check OpenAI pricing for current thresholds.
Cost controls: Batch API gives 50% off for async workloads. Cached input reduces repeat-prompt cost (e.g. long system prompts) by ~90%.
Cost vs quality framing
Not every task needs the most capable model. Frame the decision as:
- What is the cost of a wrong answer? Low (e.g. misclassified tag) vs high (e.g. code generation, legal or medical).
- How constrained is the output? Structured (JSON, fixed format) vs open-ended (prose, creative).
- Is latency critical? Real-time chat vs batch or overnight jobs.
If the cost of a wrong answer is low and the output is constrained, smaller models usually hold. If the cost of error is high or the output is open-ended, the premium tier is often justified.
When GPT-5.5 is worth it
- Complex multi-step reasoning
- Code generation with correctness requirements
- High-stakes evaluation or judgment
- Long-context analysis where retrieval is not enough
- Tasks where a cheaper model produces plausible but wrong answers
When GPT-5-mini is enough
- Classification (category, sentiment, intent)
- Entity extraction and structured output
- Simple summarization to a template
- Data normalization and formatting
- Yes/no or rule-based decisions
- Most RAG when retrieval does the heavy lifting
For a full evaluation process, see switching to cheaper AI models without losing quality.
Selection scorecard
Use this for one workflow at a time.
| Factor | GPT-5.5 | GPT-5-mini |
|---|---|---|
| Cost per 1M input | $5.00 | $0.25 |
| Cost per 1M output | $30.00 | $2.00 |
| Reasoning capability | High | Moderate |
| Instruction-following | Strong | Good for structured tasks |
| Latency | Higher (more compute) | Lower |
| Use when | Error cost is high, output is open-ended | Error cost is low, output is constrained |
Practical rule: Start with GPT-5-mini for new workflows. Escalate to GPT-5.5 only when evaluation shows the cheaper model fails your quality bar.
Migration from GPT-4o
If you are still on GPT-4o or GPT-4o-mini:
- GPT-4o was retired in February 2026; plan to move to GPT-5.5 or GPT-5-mini. Retirements like this are tracked in the monthly new and deprecated model roundup.
- GPT-5.5 is the current flagship tier (GPT-5.5 Pro for the highest-effort work).
- GPT-5-mini is the low-cost tier, comparable in role to GPT-4o-mini.
- Re-run evaluation when migrating—behavior and quality can differ even when the role is similar.
How to measure whether model choice is right
Track cost per request and cost per successful outcome. If you switch to a cheaper model:
- Compare cost per request before and after.
- Measure quality (accuracy, user feedback, error rate) on a sample.
- If quality holds, keep the cheaper model. If not, identify which task types need the premium tier.
StackSpend helps by showing spend by model, so you can see the impact of routing changes over time. See OpenAI cost monitoring.
What to do next
- Embeddings vs full context cost efficiency when retrieval strategy is in scope
- LLM cost optimization playbook
FAQ
When is the frontier OpenAI model worth the price?
When the task involves complex multi-step reasoning, long context, or high-stakes output where a wrong answer is expensive. For mechanical or well-structured work the smaller model usually matches it.
How much cheaper is the mini model?
Roughly an order of magnitude on token price. The quality gap on well-defined tasks is small enough that most teams overpay by routing everything to the larger model by default.
How do I decide which model a task needs?
Score the task on reasoning depth, context length and cost of error. High on all three justifies the frontier model; anything else should start on the smaller one and escalate only on failure.
How do I know if my model choice is right?
Measure cost per request and quality outcome per task type, not overall spend. A model change looks free in the total and obvious in the per-task number.
