GPT vs Gemini API pricing. Side by side.

OpenAI’s GPT and Google’s Gemini models, matched tier by tier, with each lab’s list price and what the same tokens cost on Zurelay, where both run on one API key.

GPT models
7
Gemini models
6
Below list
58–75%
Get your API keyCompare by tier

By tier

Matched tier by tier. As each lab describes them.

Each lab’s most capable model

Astra is the top of OpenAI’s GPT-6 family. Gemini 4 Argon is Google’s frontier model for long, complex work in software engineering, research and security.

GPT-6 Astra
OpenAI
75%off
Input / 1M
$2.50 $10.00
Output / 1M
$12.50 $50.00
Context
1.05M tokens
10K requests
$112.50
Reads
Text, images and PDFs
Gemini 4 Argon
Google
60%off
Input / 1M
$0.80 $2.00
Output / 1M
$4.00 $10.00
Context
1M tokens
10K requests
$36.00
Reads
Text, images, PDFs, audio and video

For the typical load: Gemini 4 Argon costs 68% less than GPT-6 Astra ($36.00 against $112.50).

Coding and agents

OpenAI says GPT-6.1 Sol nearly matches Astra on agentic coding. Google calls Gemini 3.8 Flash its most intelligent Flash model, built for long-horizon software engineering and agents.

GPT-6.1 Sol
OpenAI
75%off
Input / 1M
$0.50 $2.00
Output / 1M
$2.50 $10.00
Context
1.05M tokens
10K requests
$22.50
Reads
Text, images and PDFs
Gemini 3.8 Flash
Google
58%off
Input / 1M
$0.32 $0.75
Output / 1M
$1.59 $3.75
Context
1M tokens
10K requests
$14.35
Reads
Text, images, PDFs, audio and video

For the typical load: Gemini 3.8 Flash costs 36% less than GPT-6.1 Sol ($14.35 against $22.50).

Fast and low-cost

Luna is the fastest, lowest-cost GPT-6. Gemini 3.5 Flash-Lite is the fastest, most cost-effective model in Google’s Gemini 3.5 series.

GPT-6 Luna
OpenAI
75%off
Input / 1M
$0.025 $0.10
Output / 1M
$0.125 $0.50
Context
1.05M tokens
10K requests
$1.13
Reads
Text, images and PDFs
Gemini 3.5 Flash-Lite
Google
58%off
Input / 1M
$0.12 $0.30
Output / 1M
$1.04 $2.50
Context
1M tokens
10K requests
$7.60
Reads
Text, images, PDFs, audio and video

For the typical load: GPT-6 Luna costs 85% less than Gemini 3.5 Flash-Lite ($1.13 against $7.60).

The typical load is 10,000 requests of 2,000 tokens in and 500 out. Struck-through prices are each lab’s list price.

Every model

Every GPT and Gemini model. Cheapest first.

GPT and Gemini API prices on Zurelay and at each lab's list price, cheapest first
ModelContextInput / 1MOutput / 1M10K requestsYou save
GPT-6 LunaOpenAI · gpt-6-luna1.05M
$0.025
$0.10
$0.125
$0.50
$1.125
$4.50
75%off
GPT-5.6 LunaOpenAI · gpt-5.6-luna1.05M
$0.08
$0.20
$0.49
$1.20
$4.05
$10.00
59%off
Gemini 3.1 Flash-LiteGoogle · gemini-3.1-flash-lite1M
$0.10
$0.25
$0.59
$1.50
$4.95
$12.50
61%off
Gemini 3.5 Flash-LiteGoogle · gemini-3.5-flash-lite1M
$0.12
$0.30
$1.04
$2.50
$7.60
$18.50
58%off
Gemini 3.8 FlashGoogle · gemini-3.8-flash1M
$0.32
$0.75
$1.59
$3.75
$14.35
$33.75
58%off
Gemini 3.7 FlashGoogle · gemini-3.7-flash1M
$0.32
$0.75
$1.59
$3.75
$14.35
$33.75
58%off
GPT-6.1 SolOpenAI · gpt-6.1-sol1.05M
$0.50
$2.00
$2.50
$10.00
$22.50
$90.00
75%off
GPT-6 SolOpenAI · gpt-6-sol1.05M
$0.50
$2.00
$2.50
$10.00
$22.50
$90.00
75%off
GPT-5.6 TerraOpenAI · gpt-5.6-terra1.05M
$0.54
$2.00
$3.26
$12.00
$27.10
$100.00
73%off
Gemini 4 ArgonGoogle · gemini-4-argon1M
$0.80
$2.00
$4.00
$10.00
$36.00
$90.00
60%off
Gemini 3.1 ProGoogle · gemini-3.1-pro1M
$0.80
$2.00
$4.80
$12.00
$40.00
$100.00
60%off
GPT-5.6 SolOpenAI · gpt-5.6-sol1.05M
$1.14
$4.00
$5.70
$20.00
$51.30
$180.00
72%off
GPT-6 AstraOpenAI · gpt-6-astra1.05M
$2.50
$10.00
$12.50
$50.00
$112.50
$450.00
75%off

Prices in US dollars per 1M tokens, live from the catalog, with each lab's list price struck through. 10K requests is 10,000 requests of 2,000 tokens in and 500 out, at each price. You save is on output tokens.

Try both

One key for both. Test on your own work.

Benchmarks don’t tell you which model handles your prompts better. Run the same requests through both and compare the answers, the tokens and the bill.

On Zurelay, every GPT and Gemini model is on the same key and the same OpenAI-compatible endpoint, so switching is a change of model name.

compare.py
from openai import OpenAI
client = OpenAI(base_url="https://api.zurelay.com/v1", api_key="YOUR_ZURELAY_KEY")
prompt = [{"role": "user", "content": "Refactor this function: ..."}]
# Same key, same code: only the model name changes
for model in ["gpt-6.1-sol", "gemini-3.8-flash"]:
    reply = client.chat.completions.create(model=model, messages=prompt)
    print(model, reply.usage.total_tokens, reply.choices[0].message.content[:200])

FAQ

GPT vs Gemini questions.

Anything else? Ask our team and we’ll answer by email.

Is GPT or Gemini cheaper?

It depends on the tier. For 10,000 requests of 2,000 tokens in and 500 out on Zurelay: each lab's most capable model, Gemini 4 Argon costs 68% less than GPT-6 Astra ($36.00 against $112.50); coding and agents, Gemini 3.8 Flash costs 36% less than GPT-6.1 Sol ($14.35 against $22.50); fast and low-cost, GPT-6 Luna costs 85% less than Gemini 3.5 Flash-Lite ($1.13 against $7.60). Both are 58–75% below their labs' list prices.

Can I use GPT and Gemini with one API key?

Yes. One Zurelay key calls every GPT and Gemini model. With an OpenAI SDK, point it at https://api.zurelay.com/v1 and change only the model name to switch between them.

Which is better for coding?

Both labs build models for it: OpenAI says GPT-6.1 Sol nearly matches GPT-6 Astra on agentic coding and computer use, and Google built Gemini 3.8 Flash for long-horizon software engineering and autonomous agents. Gemini also reads audio and video. Results vary by task, so run your own through both. On Zurelay that's a change of model name.

Are these the real GPT and Gemini models?

Yes. Each model is served over several routes run by independent sellers, and every route is tested before it serves requests and every day after, with identity, dated-knowledge and reasoning checks that catch a different or older model sold under the name you asked for.

More pricing

Stop paying list price.
Start saving today.

Create a key in seconds, point the SDK you already use at Zurelay, and every request costs up to 90% less from the first token.

Get your API keyTalk to an engineer