Last updated: November 17, 2026 · By Visal Vadayar
Quick answer: In November 2026, mid-tier models cost about the same across providers: GPT-6.1 Sol and Claude Sonnet 5.5 are $2 input / $10 output per million tokens, and Gemini 3.1 Pro is $2 / $12. Flagships differ more: Claude Opus 5.5 is $4 / $20 while GPT-6 Astra is $10 / $50. The cheapest tiers (GPT-6 Luna, Claude Haiku 5.5, Gemini Flash-Lite) cost $0.10–$0.30 input.
Want your exact monthly bill? Plug in your requests and tokens.
Open the calculator →OpenAI API pricing
OpenAI’s current lineup is the GPT-6 family, with older GPT-5 models still available and cheap.
| Model | Input / 1M | Cached input / 1M | Output / 1M | Best for |
|---|---|---|---|---|
| GPT-6 Astra | $10.00 | $1.00 | $50.00 | Hardest reasoning and agent tasks |
| GPT-6.1 Sol | $2.00 | $0.10 | $10.00 | Default production model |
| GPT-6.1 Terra | $1.00 | $0.05 | $5.00 | Cost-sensitive production |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 | High-volume simple tasks |
| GPT-5 | $1.25 | $0.125 | $10.00 | Legacy apps |
| GPT-5 mini | $0.25 | $0.025 | $2.00 | Budget legacy |
| GPT-5 nano | $0.05 | $0.005 | $0.40 | Cheapest classification |
Watch out: on GPT-6 and GPT-5.x models, a request with more than about 272K input tokens is billed at the long-context rate (2× input, 1.5× output) for the whole request.
Claude (Anthropic) API pricing
| Model | Input / 1M | Cache read / 1M | Output / 1M | Best for |
|---|---|---|---|---|
| Claude Fable 5.2 | $8.00 | $0.20 | $40.00 | Frontier capability |
| Claude Opus 5.5 | $4.00 | $0.20 | $20.00 | Complex coding and agents |
| Claude Sonnet 5.5 | $2.00 | $0.10 | $10.00 | Everyday production work |
| Claude Haiku 5.5 (≤100K prompt) | $0.10 | $0.01 | $0.50 | Fast, cheap tasks |
| Claude Haiku 5.5 (>100K prompt) | $0.50 | $0.05 | $2.50 | Long documents |
Anthropic’s cache reads cost 5% of the normal input price on the 5.5 models, which makes Claude very cheap for apps that resend the same long system prompt.
Gemini (Google) API pricing
| Model | Input / 1M | Output / 1M | Note |
|---|---|---|---|
| Gemini 4 Argon | $2.00 | $10.00 | Rises to $4 / $20 on Dec 1, 2026 |
| Gemini 3.1 Pro | $2.00 | $12.00 | $4 / $18 above 200K tokens |
| Gemini 3.8 Flash | $0.75 | $3.75 | Rises to $1.50 / $7.50 on Jan 1, 2027 |
| Gemini 3.5 Flash-Lite | $0.30 | $2.50 | Flat rate |
| Gemini 3.1 Flash-Lite | $0.25 | $1.50 | Being replaced by 3.5 Flash-Lite |
| Gemini 2.5 Flash-Lite | $0.10 | $0.40 | Cheapest Gemini text model |
Google bills “thinking” tokens as output, and cached input costs about 90% less than fresh input. Gemini also has a free tier in Google AI Studio for testing. Learn more in our Gemini overview.
Side-by-side: same workload, three providers
Example: 50,000 requests a month, 2,000 input tokens and 500 output tokens each, no caching, standard processing.
| Tier | OpenAI | Anthropic | |
|---|---|---|---|
| Budget | GPT-6 Luna: $22.50 | Haiku 5.5: $22.50 | 2.5 Flash-Lite: $20.00 |
| Mid-tier | GPT-6.1 Sol: $450 | Sonnet 5.5: $450 | 3.1 Pro: $500 |
| Flagship | GPT-6 Astra: $2,250 | Opus 5.5: $900 | Gemini 4 Argon: $450 |
At the mid-tier, the bills are almost identical, so choose on quality, speed and features. At the flagship tier the gap is large, which is why many teams use a flagship only for the hardest 5–10% of requests.
Discounts that change the math
- Prompt caching: OpenAI GPT-6.1 Sol cached input is $0.10 vs $2.00 fresh; Claude 5.5 cache reads are 5% of input; Gemini cached input is about 10% of input.
- Batch API: all three providers charge about 50% less for asynchronous batch jobs.
- Free tiers: Google AI Studio offers free Gemini usage with rate limits, useful for prototypes.
Not sure what counts as a token? Our guide to how tokens and AI pricing work explains it with examples.
Which AI API should you choose?
- Building a chatbot on a budget: start with Claude Haiku 5.5 or GPT-6 Luna and escalate hard questions. See how much an AI chatbot costs to run.
- Coding agents and complex reasoning: Claude Opus 5.5 gives flagship-level work at a lower price than GPT-6 Astra.
- Very long documents: compare long-context thresholds carefully; Gemini Flash models have flat pricing across their 1M-token window.
- Already on Google Cloud: Gemini through Vertex AI simplifies billing and data residency.
Frequently asked questions
Is Claude cheaper than ChatGPT's API?
It depends on the tier. Claude Sonnet 5.5 and GPT-6.1 Sol have the same list price ($2 input / $10 output per million tokens). Claude Opus 5.5 ($4/$20) costs less than GPT-6 Astra ($10/$50). At the budget end, Claude Haiku 5.5 and GPT-6 Luna both cost $0.10/$0.50.
Which AI API is cheapest for high volume?
GPT-5 nano ($0.05/$0.40) and Gemini 2.5 Flash-Lite ($0.10/$0.40) have the lowest list prices, followed by GPT-6 Luna and Claude Haiku 5.5 ($0.10/$0.50). Test quality first, because a cheaper model that needs retries can cost more overall.
Do all providers offer batch discounts?
Yes. OpenAI, Anthropic and Google all offer batch processing at roughly half the standard price for work that can wait.
Why do Gemini prices change in 2027?
Google has announced that Gemini 3.8 Flash rises from $0.75/$3.75 to $1.50/$7.50 per million tokens on January 1, 2027, and Gemini 4 Argon rises from $2/$10 to $4/$20 on December 1, 2026.




