Skip to content

Pricing · Gemini

Gemini 3.1 Pro pricing: API price, cost per message and plans

What Gemini 3.1 Pro costs, three ways: Google's API price per token, what Gemini 3.1 Pro's replies actually cost in our test runs, and how many Gemini 3.1 Pro messages each llmwise plan includes. Every limit on Gemini 3.1 Pro is on its model page.

API prices checked against OpenRouter's Gemini 3.1 Pro page. Updated .

Short answer

Gemini 3.1 Pro's API price is $2.00 per million input tokens and $12.00 per million output tokens (Google's list price, September 2026; the standard rate, for prompts up to 200K tokens). In our 50 test runs a Gemini 3.1 Pro reply cost $0.0107 on average (median $0.0097), from $0.0057 a reply on agents and tool use to $0.0195 on coding. In llmwise you pay per message instead: up to 125 Gemini 3.1 Pro messages a month on Pro ($20), 400 on Max ($50), 900 on Ultra ($100) and 2,000 on Studio ($200).

Gemini 3.1 Pro API pricing

Gemini 3.1 Pro through its API is billed per token: $2.00 in, $12.00 out per million tokens, as of September 2026.

Gemini 3.1 Pro API price
Input tokens$2.00 per million
Output tokens$12.00 per million
A typical Gemini 3.1 Pro message (4,000 in, 700 out)$0.0164
1,000 typical Gemini 3.1 Pro messages$16.40
Typical Gemini 3.1 Pro messages for $160
Whose priceGoogle's list price
Notethe standard rate, for prompts up to 200K tokens

What a Gemini 3.1 Pro reply cost in our test runs

We sent the same 50 test prompts to every model in llmwise on September 27, 2026, through OpenRouter with llmwise's own prompt and settings. Here's what each Gemini 3.1 Pro reply cost, job by job.

What a Gemini 3.1 Pro reply cost in our test runs, job by job
JobGemini 3.1 Pro, per replyPer 1,000 repliesTokens in / outPassed
Coding$0.0195$19.46518 / 1,5365 of 5
Writing$0.0092$9.18427 / 6944 of 5
Math$0.0100$10.01423 / 7645 of 5
Summarization$0.0088$8.82532 / 6475 of 5
Data analysis$0.0155$15.47678 / 1,1765 of 5
Customer support$0.0106$10.58559 / 7885 of 5
Translation$0.0140$13.97467 / 1,0865 of 5
SQL$0.0077$7.72542 / 5535 of 5
RAG and answering from documents$0.0058$5.79705 / 3655 of 5
Agents and tool use$0.0057$5.68512 / 3885 of 5
Each Gemini 3.1 Pro cost is the average of what OpenRouter charged us for its replies to one job's prompts; tokens out include any reasoning. Passed: how many of those Gemini 3.1 Pro replies passed their check. In llmwise a Gemini 3.1 Pro message isn't billed by the token.

Across all 50 Gemini 3.1 Pro replies: $0.0107 on average, $0.0097 at the median, $0.0029 for the cheapest and $0.0351 for the dearest. The Gemini 3.1 Pro replies averaged 536 tokens in and 800 out, against the typical message's 4,000 in and 700 out: each test prompt starts a new chat, while a reply deep in a long chat rereads all of it.

Every prompt, and how each reply is scored.

Gemini 3.1 Pro on each llmwise plan

In llmwise a Gemini 3.1 Pro message is one of a published count, however long the reply runs, and you see what's left before you send.

  • Free: Gemini 3.1 Pro is in the one-time trial of 5 messages, no card needed.

  • Pro, $20 a month: up to 125 Gemini 3.1 Pro messages a month, which is $0.16 a Gemini 3.1 Pro message if every message goes to Gemini 3.1 Pro.

  • Max, $50 a month: up to 400 Gemini 3.1 Pro messages a month, which is $0.13 a Gemini 3.1 Pro message if every message goes to Gemini 3.1 Pro.

  • Ultra, $100 a month: up to 900 Gemini 3.1 Pro messages a month, which is $0.11 a Gemini 3.1 Pro message if every message goes to Gemini 3.1 Pro.

  • Studio, $200 a month: up to 2,000 Gemini 3.1 Pro messages a month, which is $0.10 a Gemini 3.1 Pro message if every message goes to Gemini 3.1 Pro.

  • A $10 top-up adds up to 100 Gemini 3.1 Pro messages, valid 12 months.

Every limit is published. Paid plans also have a monthly fair-use limit on AI cost: Pro $7.50, Max $20, Ultra $42, Studio $85. Using every message on your plan at typical sizes stays under it; very large messages and heavy research use it faster. Every limit, explained.

Gemini 3.1 Pro against the other Gemini models

Every Gemini model in llmwise beside Gemini 3.1 Pro, dearest first.

Gemini 3.1 Pro beside its family
ModelAPI, per 1M in / outInput price against Gemini 3.1 ProAverage reply in our runsOn Pro
Gemini 3.1 Pro (preview)GoogleThis page$2.00 / $12.00–$0.0107Up to 125 a month
Gemini 3.8 FlashGoogle$0.75 / $3.7538% of Gemini 3.1 Pro's$0.0013Up to 250 a month
API prices as of September 2026, from our model catalog. The average reply is what OpenRouter charged for each model's replies to the same test prompts as Gemini 3.1 Pro's. On Pro: the most messages a month if all of Pro's allowance goes to that model, or an everyday model's daily count.

Gemini 3.1 Pro in Google's own plans

What Google's own subscriptions say about Gemini 3.1 Pro, in their words, set beside llmwise's count. Each was checked on the day shown; we drop a plan from this page when its facts are more than 45 days old.

Gemini 3.1 Pro's own page: Gemini 3.1 Pro: price, limits and messages on every plan. Gemini 3.1 Pro passed 49 of the 50 test prompts these costs come from.

Questions

How much does a Gemini 3.1 Pro message cost?

In our test runs, $0.0107 a reply on average through OpenRouter (median $0.0097), across 50 Gemini 3.1 Pro replies to our test prompts. At the API price a typical Gemini 3.1 Pro message of 4,000 tokens in and 700 out costs $0.0164; a long chat costs more, because every reply rereads it.

What is Gemini 3.1 Pro's API price?

$2.00 per million input tokens and $12.00 per million output tokens, as of September 2026 (Google's list price).

How many Gemini 3.1 Pro messages does llmwise Pro include?

Up to 125 Gemini 3.1 Pro messages a month on Pro ($20 a month), if the whole allowance goes to Gemini 3.1 Pro; the count on each plan is above.

Can I use Gemini 3.1 Pro for free?

Yes, to try: Gemini 3.1 Pro is in llmwise's one-time free trial of 5 messages.

Claude, GPT, Gemini, DeepSeek, Grok, Kimi, and GLM, in one chat.

See what a message costs before you send it. Free is 5 messages to try; sign in with an email link, no password or card.