OpenAI vs Anthropic API pricing
Updated 2026-10-18. Prices from official pricing pages, quality ranks from LMArena.
At the top end they cost the same: GPT-6 Astra and Claude Fable 5.1 are both $10 input and $50 output per 1M tokens. In the mid tier, GPT-6.1 Sol ($2 / $10) and Claude Sonnet 5.5 ($2 / $10) also match. OpenAI is far cheaper at the small end: GPT-6 Luna is $0.10 / $0.50 vs Claude Haiku 5.5 at $0.50 / $2.50. Claude models rank higher on LMArena's Code Arena among the models we track.
| Model | Input | Output | 30M in + 6M out |
|---|---|---|---|
| GPT-6 Astra | $10 | $50 | $600 |
| GPT-5.5 | $5 | $30 | $330 |
| GPT-5.6 Sol | $4 | $20 | $240 |
| GPT-5.6 Terra | $2 | $12 | $132 |
| GPT-6.1 Sol | $2 | $10 | $120 |
| GPT-6 Sol | $2 | $10 | $120 |
| GPT-6 Luna | $0.10 | $0.50 | $6 |
| Claude Fable 5.1 | $10 | $50 | $600 |
| Claude Opus 5.5 | $4 | $20 | $240 |
| Claude Sonnet 5.5 | $2 | $10 | $120 |
| Claude Haiku 5.5 | $0.50 | $2.50 | $30 |
USD per 1M tokens, standard tier, short-context rate, checked 2026-10-18. Coding rank is the position among models we track on LMArena (CC BY 4.0) Code Arena.
Tier by tier
- Flagship: GPT-6 Astra $10 / $50 vs Claude Fable 5.1 $10 / $50. Same list price; Astra charges more above 272K prompt tokens (Longer prompts cost $20 in, $75 out).
- Workhorse: GPT-6.1 Sol $2 / $10 vs Claude Sonnet 5.5 $2 / $10. Sample month: $120 vs $120. Sol has cheaper cache reads ($0.10).
- Top coding model: Claude Opus 5.5 at $4 / $20 is #1 on Code Arena among tracked models, ahead of GPT-6 Astra (#2) at a lower price.
- Small and cheap: GPT-6 Luna $0.10 / $0.50 vs Claude Haiku 5.5 $0.50 / $2.50. Sample month: $6 vs $30, about 5x apart.
Things the headline price hides
- Long prompts: OpenAI charges a higher rate above 272K prompt tokens on GPT-6 models. Claude Haiku 5.5's listed rate is for prompts up to 100K tokens.
- Caching: both discount cached input heavily; Anthropic also bills cache writes (for example $2.50 per 1M for 5-minute writes on Sonnet 5.5).
- Subscriptions with API credits: Claude Max 5x and Max 20x include $100/mo API credits and $200/mo API credits respectively, per Anthropic's pricing page.
Other providers are often cheaper still: see the cheapest LLM API ranking, DeepSeek vs Claude API cost, or the full AI API price table.
Sources
FAQ
Is the OpenAI API cheaper than Claude?
At the small end, yes: GPT-6 Luna costs $0.10 input and $0.50 output per 1M tokens vs $0.50 and $2.50 for Claude Haiku 5.5. Flagship and mid-tier list prices are the same.
How much does GPT-6.1 Sol cost vs Claude Sonnet 5.5?
Both list at $2 input and $10 output per 1M tokens. GPT-6.1 Sol has cached input at $0.10.
Which is better for coding, GPT or Claude?
On LMArena's Code Arena, Claude Opus 5.5 is #1 among the models we track, GPT-6 Astra #2, Claude Sonnet 5.5 #3 and GPT-6.1 Sol #4. Ranks shift, so test on your own tasks.
What is the cheapest OpenAI model?
GPT-6 Luna at $0.10 input and $0.50 output per 1M tokens, as of 2026-10-18.
What is the cheapest Claude model?
Claude Haiku 5.5 at $0.50 input and $2.50 output per 1M tokens, as of 2026-10-18.
Is a ChatGPT or Claude subscription cheaper than the API?
For steady daily use it often is, because plans include a usage allowance for a flat price. The plan vs API calculator compares both for your usage.
Related
- OpenAI API pricing
- Anthropic API pricing
- Cheapest LLM API
- DeepSeek vs Claude API cost
- Best model for coding
- Plan vs API calculator
More guides
- Grok limits explained
- Grok usage limit reached: what to do
- How to check your remaining AI usage
- Is ChatGPT Plus worth it?
- Is Claude Max worth it?
- Is Cursor worth it?
Not sure a plan beats the API for you? Try the plan vs API cost calculator.