Gemini 3.8 Flash pricing on CheapRelay: $0.95 input per 1M

01 / Gemini 3.8 Flash pricing on CheapRelay, line by line

What does Gemini 3.8 Flash cost on CheapRelay? On the current price list, the model id is gemini-3.8-flash, and the price is $0.95 per 1,000,000 input tokens, $4.28 per 1,000,000 output tokens, and $0.0944 per 1,000,000 cache-read tokens. That is the complete Gemini 3.8 Flash pricing picture: three token lines, no subscription requirement, and no monthly fee.

Why show these numbers first? Because a token price only matters when you can see how it becomes a bill. CheapRelay charges for real tokens against a prepaid balance, so the useful question is not whether a headline rate looks attractive. It is what a particular request will cost after input, output, and cache reads are added up.

• ■ Input: $0.95 per 1,000,000 tokens.
• ■ Output: $4.28 per 1,000,000 tokens.
• ■ Cache read: $0.0944 per 1,000,000 tokens.
• ■ Model id to enter: gemini-3.8-flash.

There is no official list price for Gemini 3.8 Flash on this price list, so this article does not attach one to it. Instead, it compares Gemini 3.8 Flash with the other Gemini models that do appear on the CheapRelay price list, using the published figures directly.

02 / How much does one Gemini 3.8 Flash request cost?

How much is a full 1M-token example? Example: one request uses 1,000,000 input tokens, 1,000,000 output tokens, and 1,000,000 cache-read tokens. The arithmetic is $0.95 + $4.28 + $0.0944 = $5.3244. That is the example total for those three lines at the published Gemini 3.8 Flash rates.

What about a more typical mixed workload? Example: a request uses 100,000 input tokens, 20,000 output tokens, and 50,000 cache-read tokens. The input line is 100,000 divided by 1,000,000, multiplied by $0.95, which equals $0.095. The output line is 20,000 divided by 1,000,000, multiplied by $4.28, which equals $0.0856. The cache-read line is 50,000 divided by 1,000,000, multiplied by $0.0944, which equals $0.00472. Added together, the example cost is $0.095 + $0.0856 + $0.00472 = $0.18532.

Why is output the line to watch? The published output rate is $4.28 per 1M tokens, while the input rate is $0.95 per 1M. In the example above, 100,000 input tokens cost $0.095, but 20,000 output tokens cost $0.0856. The smaller output quantity is close to the larger input cost because the per-token output rate is higher. A developer who controls prompt length and asks for concise answers can therefore see exactly where the budget goes.

What does the cache-read line do to the arithmetic? Example: 1,000,000 cache-read tokens cost $0.0944 on Gemini 3.8 Flash. That is $0.0944 compared with $0.95 for 1,000,000 input tokens and $4.28 for 1,000,000 output tokens. In this comparison, cache reads are the smaller line of the three. If your workflow reuses a large fixed context, that line is worth tracking instead of guessing.

Can the sign-up credit cover a real example? Every new CheapRelay account receives $0.09 in sign-up credit, usable on any model. Example: the mixed request above costs $0.18532, so the $0.09 credit covers $0.09 divided by $0.18532, or about 48.57% of that example request. Example: measured only against input, $0.09 divided by $0.95 covers about 94,737 input tokens. The credit is small, but it is usable immediately rather than being tied to one model.

03 / How does Gemini 3.8 Flash compare with other Gemini models?

How does Gemini 3.8 Flash compare with Gemini 2.5 Flash? Gemini 2.5 Flash is listed at $0.1778 input, $1.42 output, and $0.0181 cache read per 1M tokens. Gemini 3.8 Flash is listed at $0.95 input, $4.28 output, and $0.0944 cache read. The input figure is $0.95 against $0.1778, the output figure is $4.28 against $1.42, and the cache-read figure is $0.0944 against $0.0181. On all three lines, Gemini 3.8 Flash has the higher rate, so a request that uses the same token counts will cost more on Gemini 3.8 Flash.

By how much? Example: 1M input, 1M output, and 1M cache read cost $0.1778 + $1.42 + $0.0181 = $1.6159 on Gemini 2.5 Flash. The same token counts cost $5.3244 on Gemini 3.8 Flash. The difference is $5.3244 - $1.6159 = $3.7085, and $5.3244 divided by $1.6159 is about 3.30 times as much. That gap is a feature, not a problem: the price list lets you decide which Gemini model fits the job and budget instead of paying one blended rate for everything.

How does it compare with Gemini 3.1 Pro? Gemini 3.1 Pro is listed at $1.12 input, $6.49 output, and $0.1125 cache read per 1M tokens. Gemini 3.8 Flash is listed at $0.95 input, $4.28 output, and $0.0944 cache read. Here, Gemini 3.8 Flash has the lower figure on each line. Example: the same 1M/1M/1M mix costs $1.12 + $6.49 + $0.1125 = $7.7225 on Gemini 3.1 Pro and $5.3244 on Gemini 3.8 Flash. The example saving is $7.7225 - $5.3244 = $2.3981.

So which one should you pick? If output is the dominant cost, Gemini 2.5 Flash has the smaller output rate at $1.42 compared with $4.28. If you want a step up from Gemini 2.5 Flash but want to keep the input rate below Gemini 3.1 Pro's $1.12, Gemini 3.8 Flash sits at $0.95. Those are the numbers. CheapRelay puts all three on the price list so the choice can be made before the request runs.

04 / How do I use Gemini 3.8 Flash on CheapRelay?

Which client should I use? Gemini models use the native protocol only, under /v1beta, through Gemini CLI. Point Gemini CLI at the CheapRelay API base, https://api.cheapapirelay.com, and enter gemini-3.8-flash as the model id. That is the supported route for this model.

Can I use an OpenAI-compatible client? Not for Gemini on CheapRelay. Gemini is native-protocol only through Gemini CLI and is not served through the OpenAI-compatible route. This matters because copying an OpenAI-compatible configuration into Gemini CLI will not match how the model is exposed.

What does the setup involve? Example: create a CheapRelay account, use the $0.09 sign-up credit or add prepaid funds, configure Gemini CLI with the API base and the native /v1beta path, then select gemini-3.8-flash. There is no monthly fee and no package to buy. The account uses a prepaid balance, and the balance does not expire, so unused funds can stay on the account rather than being tied to a recurring billing cycle.

Where do I check the current prices? The current price list at CheapRelay is the reference for Gemini 3.8 Flash and every other listed model. The Gemini 3.8 Flash line shows $0.95 input, $4.28 output, and $0.0944 cache read per 1M tokens. Check that list when you set a budget, because the arithmetic above depends on those published rates.

05 / How does prepaid billing and account credit work?

How is CheapRelay billed? You pay per real token from a prepaid balance. There is no monthly fee. The balance does not expire, which means a small balance can remain available for the next request instead of requiring a fresh recurring purchase.

What happens if the balance runs out? HTTP 402 means the balance is at or below zero. That is a clear signal to top up before continuing, rather than discovering the problem from a failed workflow halfway through a task.

Can I get unused funds back? Yes, within the stated refund window. Unused balance can be refunded within 12 days of the top-up, minus what was spent and the payment fee. Example: if a top-up includes funds you never used, the refund calculation starts with the unused balance and subtracts the spending and the payment fee. Read those two deductions before adding money you may need back quickly.

Can referrals reduce the effective cost? The referral program pays 20% of what a directly invited user pays, plus 6% from the second level. Those amounts go into a withdrawable rewards balance. The smallest withdrawal is ¥20, via Alipay. This is separate from your prepaid API balance: referral rewards are withdrawable rewards, while API usage is paid from the prepaid balance.

06 / What limits and terms apply to Gemini 3.8 Flash on CheapRelay?

How many calls can I make? Each key is limited to 120 calls per minute, and each account can run 8 calls at once. Example: if you are running a batch job, keep the per-key request rate under 120 calls per minute and the account's simultaneous calls at or below 8. Those limits apply across the account setup rather than to a single model, so they are worth checking before you launch a burst of Gemini 3.8 Flash requests.

What is the final cost check before I start? Example: a workload with 100,000 input tokens, 20,000 output tokens, and 50,000 cache-read tokens costs $0.18532 on Gemini 3.8 Flash at $0.95, $4.28, and $0.0944 per 1M tokens. Example: a 1M/1M/1M mix costs $5.3244. Example: the same 1M/1M/1M mix costs $7.7225 on Gemini 3.1 Pro and $1.6159 on Gemini 2.5 Flash.

Why does that arithmetic matter? It turns model selection into a budget decision. Gemini 2.5 Flash costs less per token on all three lines. Gemini 3.8 Flash costs less per token than Gemini 3.1 Pro on all three lines. CheapRelay gives you the published rates, the native Gemini CLI route, prepaid billing with no monthly fee, and the arithmetic needed to check the bill before it grows.