GPT-5.5 Pricing: How Much Does It Cost in 2026?
OpenAI released GPT-5.5 on April 23, 2026, and with it came a detailed pricing structure across multiple tiers. Whether you're accessing it through ChatGPT, Codex, or the API, here's everything you need to know about GPT-5.5's cost.
GPT-5.5 API Pricing
API access became available on April 24, 2026 — one day after ChatGPT rollout. The pricing is:
| Model | Input (per 1M tokens) | Output (per 1M tokens) |
|---|---|---|
| gpt-5.5 | $5.00 | $30.00 |
| gpt-5.5-pro | $30.00 | $180.00 |
Special Rate Tiers
| Processing Mode | Rate |
|---|---|
| Batch / Flex pricing | 50% of standard rate |
| Priority processing | 2.5× standard rate |
Batch/Flex is ideal for non-time-sensitive workloads where you can tolerate longer latency. Priority processing is for workloads that need guaranteed fast response times ahead of standard queue traffic.
Is GPT-5.5 More Expensive Than GPT-5.4?
Yes — but it's also more token-efficient. OpenAI notes that GPT-5.5 completes the same Codex tasks with fewer tokens and fewer retries than GPT-5.4. The result for many production workflows is that the higher per-token price is partially offset by reduced total token consumption.
For Codex-heavy teams, OpenAI specifically tuned GPT-5.5 to deliver better results with fewer tokens at subscription levels.
GPT-5.5 ChatGPT Subscription Access
GPT-5.5 is included in ChatGPT paid plans — no separate API billing required for end users:
| Plan | GPT-5.5 Access | GPT-5.5 Pro Access |
|---|---|---|
| Free | ❌ | ❌ |
| Plus | ✅ | ❌ |
| Pro | ✅ | ✅ |
| Business | ✅ | ✅ |
| Enterprise | ✅ | ✅ |
Free-tier users did not receive GPT-5.5 access at launch.
GPT-5.5 in Codex: Pricing
For Codex (OpenAI's agentic coding environment), GPT-5.5 is available to:
- Plus, Pro, Business, Enterprise, Edu, and Go plan users
Codex context window: 400,000 tokens
GPT-5.5 in Codex is also available in Fast Mode, which generates tokens 1.5× faster at 2.5× the standard cost — useful for latency-sensitive automated coding pipelines.
How Does GPT-5.5 Pro Justify Its $30 Input / $180 Output Price?
GPT-5.5 Pro is aimed at the highest-accuracy use cases. Its benchmark improvements over base GPT-5.5:
| Benchmark | Base GPT-5.5 | GPT-5.5 Pro |
|---|---|---|
| BrowseComp | 84.4% | 90.1% |
| FrontierMath Tier 4 | 35.4% | 39.6% |
| GeneBench | 25.0% | 33.2% |
| FrontierMath Tier 1–3 | 51.7% | 52.4% |
Early testers described GPT-5.5 Pro responses as "significantly more comprehensive, well-structured, accurate, relevant, and useful" compared to GPT-5.4 Pro — with especially strong performance in business, legal, education, and data science.
For high-stakes professional tasks where accuracy is more important than cost, GPT-5.5 Pro delivers meaningful quality gains.
Cost Estimation: Real-World Scenarios
Scenario 1: Content Generation Pipeline
- Average article: ~3,000 input tokens + 2,000 output tokens
- Cost per article: (3,000 × $5 + 2,000 × $30) / 1,000,000 = $0.075
- 10,000 articles: ~$750
Scenario 2: Code Review System
- Average review: ~20,000 input tokens + 5,000 output tokens
- Cost per review: (20,000 × $5 + 5,000 × $30) / 1,000,000 = $0.25
- 1,000 reviews/month: ~$250
Scenario 3: Document Analysis
- 100K-token document + 5,000-token summary
- Cost: (100,000 × $5 + 5,000 × $30) / 1,000,000 = $0.65
Getting the Most Value from GPT-5.5
To maximize cost efficiency:
- Use Batch/Flex for non-urgent workloads (50% savings)
- Use gpt-5.5 (base) for the majority of tasks — reserve Pro for highest-stakes queries
- Leverage token efficiency — GPT-5.5's improved instruction following means fewer iterations
- Use platforms with built-in workflows like Framia.pro, which provides GPT-5.5-powered tools without requiring custom API integration or per-token billing management
Summary
- API base price: $5 input / $30 output per 1M tokens
- API Pro price: $30 input / $180 output per 1M tokens
- Batch/Flex: 50% discount
- Priority: 2.5× premium
- ChatGPT: Included in Plus, Pro, Business, Enterprise plans
- Context window: 1M tokens (API), 400K (Codex)
- Token efficiency: Better than GPT-5.4 — fewer tokens needed per task