Skip to content

Pricing

Pay as you go.

No subscription and no plan to pick. Add credits and spend them on any model – in chat, code, image, video, music, consensus and the API. Each request costs the model's price. If a model's usual provider is unavailable and another serves your request, the rate for that request may be higher. If the usage the model reports for a request comes to more than the published rates give, the charge may follow that report; the balance never becomes negative.

How credits work

  • 1 credit = $1

    Credits cost their face value in US dollars. VAT is added at checkout where it applies.

  • Top up 10 to 2,000

    Add any whole amount from 10 to 2,000 credits. You pay their face value plus VAT, and checkout shows the VAT and the total before you pay.

  • 12 months

    Every top-up lasts 12 months from the day you buy it. The credits that expire first are spent first.

  • Per request

    Each request is charged at the model's price below. If a model's usual provider is unavailable and another serves your request, the rate for that request may be higher. You see an estimate before a run and the exact cost after it.

What a top-up costs

The amounts in the top-up dialog, before VAT, and in brackets with the 24% VAT a consumer in the EU pays. You can also enter any other amount from 10 to 2,000 credits.
CreditsYou pay
10 credits$10 + VAT ($12.40 incl. 24% EU VAT)
25 credits$25 + VAT ($31 incl. 24% EU VAT)
50 credits$50 + VAT ($62 incl. 24% EU VAT)
100 credits$100 + VAT ($124 incl. 24% EU VAT)

At checkout the VAT follows your billing country and is shown before you pay. A consumer anywhere in the EU pays Estonia's 24% VAT, as ewpire is based in Estonia; outside the EU, checkout adds no VAT. On a business account with a valid VAT ID from an EU country other than Estonia, no VAT is added: the reverse charge applies.

What a request costs

Examples counted from the prices in the tables below. The exact charge follows the usage the model reports for your request.
  • Claude Opus 5.5: a request with 2,000 input tokens and 1,000 output tokens costs 0.02716 credits – 0.00776 for the input and 0.0194 for the output.
  • Nano Banana 2.1: 10 images cost 0.33029 credits, at 0.033029 per image.
  • Gemini Omni Flash 1.2: 5 seconds of video cost 0.64706 credits, at 0.129412 per second.

Tokens a model spends on reasoning count as output tokens.

Model prices

In credits, from the model registry: the rates a request is charged at, rounded in these tables – each model's page gives them in full. 1 credit = $1.

Text models, per million tokens

ModelInput / 1M tokensOutput / 1M tokens
Claude Opus 5.53.8819.4
GPT-6.2 Sol272,000+ input tokens: 4.964706 in · 18.617648 out2.482412.4118
GPT-6.1 Sol272,000+ input tokens: 3.88 in · 14.55 out1.949.7
GPT-6 Sol272,000+ input tokens: 3.88 in · 14.55 out1.949.7
Claude Sonnet 5.51.949.7
Claude Sonnet 51.949.7
Gemini 4 Argon1.949.7
Gemini 3.8 Flash0.80034.0013
GPT-6.1 Terra272,000+ input tokens: 1.94 in · 7.275 out0.974.85
Claude Haiku 5.5100,000+ input tokens: 0.485 in · 2.425 out0.0970.485
GPT-6 Luna272,000+ input tokens: 0.194 in · 0.7275 out0.0970.485
DeepSeek V4.1 Flash0.06330.7037
DeepSeek V4 Pro1.11711.8618
Kimi K3.1 Pro0.298310.4287
Claude Opus 54.8524.25
Grok 4.7200,000+ input tokens: 3.88 in · 11.64 out1.945.82
Kimi K30.238319.8588
Qwen4 Max1.76875.306
Qwen3.8 Max2.48247.4471
GPT-6.2 Astra272,000+ input tokens: 19.858824 in · 74.470589 out9.929449.6471
GPT-6 Astra272,000+ input tokens: 19.4 in · 72.75 out9.748.5
GLM 5.30.08681.7376
Mistral Large 40.86882.6065
Mistral Medium 3.51.86189.3088
Claude Fable 5.29.929449.6471
Claude Fable 5.112.411862.0588
Gemini 3.6 Flash0.72753.6375
GPT-5.6 Sol272,000+ input tokens: 7.76 in · 29.1 out3.8819.4
Qwen3.8 Flash0.18620.5834
Kimi K2.7 Code0.86884.6544
Qwen3.7 Plus256,000+ input tokens: 1.19153 in · 4.766118 out0.39721.5887
GPT-5.5272,000+ input tokens: 9.7 in · 43.65 out4.8529.1
Mistral Large 30.62061.8618
Claude Opus 4.74.8524.25
Devstral 20.49652.4824
Gemini 3.5 Flash Lite0.37243.1029
GLM 51.24123.9718
Gemini 2.5 Pro200,000+ input tokens: 3.102942 in · 18.617648 out1.551512.4118
Llama 4 Maverick0.43441.2412
DeepSeek V30.31951.2768
Qwen-Plus256,000+ input tokens: 0.968118 in · 2.904353 out0.32270.9681
Kimi K20.70752.8547
Mistral Large2.48247.4471
Muse Spark 1.31.55155.275
Command A+0.37241.8618
DeepSeek V3.20.32270.4716
DeepSeek V4 Flash Vision Exp0.54611.6384
Gemini 3.1 Pro Preview200,000+ input tokens: 4.964706 in · 22.341177 out2.482414.8941
Gemini 3.1 Pro Preview Custom Tools200,000+ input tokens: 4.964706 in · 22.341177 out2.482414.8941
Gemini 4 Carbon2.482412.4118
Gemini 4 Hydrogen1.86189.3088
Gemma 4 26B A4B 0.0950.3165
Gemma 4 31B0.24820.4965
GLM 4.6V0.37241.1171
GLM 5 Turbo1.48944.9647
GLM 5.3 Flash0.04950.0973
GLM 5.3 FlashX0.45921.5515
GLM 5.3 FlashX Legal0.45921.5515
GLM 5.3 Prime2.08526.5534
GLM 5V Turbo1.48944.9647
GPT Chat Latest6.205937.2353
GPT-5.1-Codex-Max1.551512.4118
GPT-5.3-Codex2.172117.3765
GPT-5.4 Mini0.93095.5853
GPT-5.4 Nano0.24821.5515
GPT-5.5 Pro272,000+ input tokens: 74.470589 in · 335.117648 out37.2353223.4118
MiniMax M30.37241.4894
MiniMax M3.1 Flash0.18620.7447
Muse Ember0.12410.6206
Muse Glimmer 30B0.43441.8618
Muse Spark Open1.55155.275
Muse Wonder4.964724.8235
Nemotron 3 Nano 30B A3B0.07450.2979
Nemotron 3 Super0.09310.4965
Nemotron 3 Ultra0.62062.7306
Nemotron 3.5 Lightning0.08690.2482
Nemotron 3.5 Super VL0.18620.6206
Nova 2 Lite0.37243.1029
Qwen3 Coder Next0.14890.9929
Qwen3.5 397B A17B0.55853.7235
Qwen3.5-9B0.16760.2458
Qwen3.6 35B A3B0.12411.1171
Qwen3.8 2.4T A95B2.48247.4471
Qwen3.8 27B0.07452.7306
Qwen3.8 Omni Flash0.18620.5834
Seed 2.1 Turbo0.62063.1029
Seed-2.0-Lite128,000+ input tokens: 0.620589 in · 4.964706 out0.31032.4824
Seed-2.0-Mini128,000+ input tokens: 0.248236 in · 0.992942 out0.12410.4965
Sonar+ 0.0062 credits / request1.24121.2412

Image models

ModelPer imageInput / 1M tokensOutput / 1M tokens
Nano Banana 2.10.033––
Nano Banana 2 (Gemini 3.1 Flash Image)0.0653––
Nano Banana Pro (Gemini 3 Pro Image)0.131––
GPT-5.4 Image 20.15799.929418.6176
Nano Banana (Gemini 2.5 Flash Image)0.0376––
Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)0.0327––
GPT-5 Image Mini0.04223.10292.4824
GPT-5 Image0.210312.411812.4118
GPT Image 2.50.0353––

Video models, per second

ModelPer second
Gemini Omni Flash 1.20.1294
Gemini Omni Flash+ 0.1235 credits / request0.0618
Kling 3.0 Turbo0.1482
Kling 4.00.25
Seedance 2 Mini0.0482
Seedance 2.50.9294
Veo 3.1 Fast0.1324
Wan 3.0 Video0.1882
Wan 3.0 Video Prime0.2965

Music models, per song

ModelPer song
Lyria 3 Pro0.0993
Lyria 3 Clip0.0496
Suno V60.0706
Suno V6 Mini0.0706
Suno V6 Wild0.0706

Where a model lists a rate for long requests, the whole request is charged at that rate once its input reaches the size shown.

Compare every model

The account is free

Log in, look around the workspace and set up projects without paying. A new account gets 1 free credit, valid for 14 days, for the starter models; it is not available through the API. After that a run needs credits: from $10 plus VAT.

Team workspaces are coming soon

One shared balance for everyone in a workspace, with a monthly limit for each member.

Pricing questions

Are there free credits or a free trial?
The account is free, and so is looking around the workspace. A new account gets 1 free credit, valid for 14 days, for the starter models; it is not available through the API. Every other model opens with a top-up, from 10 credits.
Is there anything to pay on top of the credits?
Only VAT, where it applies. Top up 100 credits and you pay $100 plus VAT – $124 for a consumer in the EU – and 100 credits land on your balance. Checkout shows the VAT and the total before you pay.
Do credits expire?
Yes, 12 months after the top-up they came from. The credits that expire first are spent first, and Billing shows each top-up with its expiry date.
How do I know what a request will cost?
Before a run ewpire shows an estimate in credits; after it, the exact charge appears in Usage. Through the API, every answer carries what it cost and your balance. A request that fails is not charged, even if part of it was already shown, and neither is one you cancel before anything is delivered. Work you stop is charged for what was delivered: a chat answer you stop, for the part already written; a Consensus you stop, or whose synthesis fails, once two or more models have answered, for the model calls already completed – with fewer than two answers it is not charged. A Code run whose answer arrives in full is charged for it, even if no new version can be saved from it. A video or song already sent to the model can't be cancelled or deleted until it finishes, and is charged only if it completes.
Can I get a refund?
You can withdraw from a top-up within 14 days of buying it, as long as you have not used any credit from it. Once you use a credit from it, your right to withdraw from that top-up ends, as you agreed at checkout. Withdraw with "Withdraw from contract here" in the footer – no sign-in needed – or in Billing, and you get the whole payment back. Top-ups on a business account have no 14-day withdrawal.
Which currency do I pay in?
US dollars, wherever you are: 1 credit is $1, plus VAT.
Is VAT included?
Prices are shown before VAT, and in brackets with the 24% VAT a consumer in the EU pays. At checkout the VAT follows your billing country and is shown before you pay: a consumer anywhere in the EU pays Estonia's 24% VAT, as ewpire is based in Estonia, and outside the EU checkout adds no VAT. On a business account with a valid VAT ID from an EU country other than Estonia, no VAT is added: the reverse charge applies.
Is there a subscription or a monthly plan?
No. ewpire has no subscription, no plan to pick and no monthly minimum: you add credits when you want, and nothing renews.

Pay as you go.

Every top-up lasts 12 months from the day you buy it. The credits that expire first are spent first.