Trusted by 332+ customers
Claude. GPT. Gemini.One plan.
Top AI models in your terminal. Pick a plan, paste the key, code. Prepaid in PLN, no subscription.
Requests through us
—
today —
USD processed
—
today —
- ——
- ——
- ——
- ——
—
Tokens
- 12 models
- 3 vendors, one key
- Ultra every 14 days
- separate limit for model issues
- 0 subscriptions
- prepaid, no surprises
- No auto-renewal
- Prepaid
- PL / EN
- Discord login
Where you code
Code where you like.
EpsilonCLI is optional. Epsilon works in Claude Code: in the terminal, in VS Code and in the desktop app. Paste the key and code like you always do.
FULL SUPPORT▐▛███▜▌ ▝▜█████▛▘ ▘▘ ▝▝
Claude Code v2.1.312
Opus 5.5 with medium effort · API Usage Billing
~\my-app
⏵⏵ auto mode on (shift+tab to cycle)
How requests work
Two limits: Normal and Ultra.
Your monthly dollars
Every request costs its API price. Resets on the 1st of every month.
Only when something goes wrong
Secure Guard block (Anthropic charges for it even with no output), a model error or an attempt to manipulate the model. Renews every 14 days.
Why Epsilon
More limit for less.
How many dollars of API usage you get for your money.
Epsilon Pro
Best value$6 000
limit per month
worth ~23 600 zł
You pay PLN 80.00
Official Claude Code Pro
42× less~$180
limit per month · estimate
You pay PLN 100.00
Another Claude provider
10× less$3 000
limit per month
You pay $100 (~PLN 393.00)
Another provider
24× less$1 208
limit one time
You pay $100 (~PLN 393.00)
Real models
Straight from the vendor's API. Not scraped from Claude Code or Kiro, like most providers do.
Fast
No shared scraped accounts and no hidden throttling. Answers come at full speed.
Quality first
We would rather be good than the cheapest. Stable answers matter more to us than shortcuts.
Competitor prices from their public offers, USD at ~3.93 PLN. Claude Code Pro usage is an estimate based on ~$6 a day of average API-equivalent usage.
Models
12 models. 3 vendors.
12 online now
Anthropic8
Claude Fable 5.1
claude-fable-5.1
- Context
- 200k
- Min. plan
- Max x3
Claude Fable 5
claude-fable-5
- Context
- 200k
- Min. plan
- Max x3
Claude Opus 5.5
claude-opus-5.5
- Context
- 200k
- Min. plan
- Pro
Claude Opus 5
claude-opus-5
- Context
- 200k
- Min. plan
- Pro
Claude Opus 4.8
claude-opus-4.8
- Context
- 200k
- Min. plan
- Pro
Claude Sonnet 5.5
claude-sonnet-5.5
- Context
- 200k
- Min. plan
- Lite
Claude Haiku 5.5
claude-haiku-5.5
- Context
- 200k
- Min. plan
- Lite
Claude Sonnet 5.5 Slow
claude-sonnet-5.5-slow
- Context
- 200k
- Min. plan
- Pro
OpenAI3
Model list can change day to day.
Pricing
Pay once. Use all month.
Lite
Without Opus
PLN 40.00
/ 30 days
Monthly limit
$1 000
≈ 3 900 PLN at API prices
Ultra limit
450 Ultra credits / 14 days
What you get
- Claude Sonnet, GPT, Gemini
- 1M context
- Top queue priority
Models
- Claude Fable 5.1
- Claude Fable 5
- Claude Opus 5.5
- Claude Opus 5
- Claude Opus 4.8
- Claude Sonnet 5.5
- Claude Haiku 5.5
- GPT-6 Astra
- GPT-6.1 Sol
- Gemini 4 argon
- Claude Sonnet 5.5 Slow
- GPT-6 Astra Slow
Pro
Opus, 1M context, slow Sonnet
PLN 80.00
/ 30 days
Monthly limit
$6 000
≈ 23 600 PLN at API prices
Bigger Ultra limit
900 Ultra credits / 14 days
What you get
- Claude Sonnet, GPT, Gemini
- 1M context
- Top queue priority
Models
- Claude Fable 5.1
- Claude Fable 5
- Claude Opus 5.5
- Claude Opus 5
- Claude Opus 4.8
- Claude Sonnet 5.5
- Claude Haiku 5.5
- GPT-6 Astra
- GPT-6.1 Sol
- Gemini 4 argon
- Claude Sonnet 5.5 Slow
- GPT-6 Astra Slow
Max x3
All models, incl. Fable, slow Sonnet and slow GPT-6 Astra
PLN 140.00
/ 30 days
Monthly limit
$15 000
≈ 59 000 PLN at API prices
Big Ultra limit
1 500 Ultra credits / 14 days
What you get
- Claude Sonnet, GPT, Gemini
- 1M context
- Top queue priority
Models
- Claude Fable 5.1
- Claude Fable 5
- Claude Opus 5.5
- Claude Opus 5
- Claude Opus 4.8
- Claude Sonnet 5.5
- Claude Haiku 5.5
- GPT-6 Astra
- GPT-6.1 Sol
- Gemini 4 argon
- Claude Sonnet 5.5 Slow
- GPT-6 Astra Slow
Max x5
PopularSame as Max x3, top queue priority
PLN 220.00
/ 30 days
Monthly limit
$30 000
≈ 117 900 PLN at API prices
Biggest Ultra limit
2 200 Ultra credits / 14 days
What you get
- Claude Sonnet, GPT, Gemini
- 1M context
- Top queue priority
Models
- Claude Fable 5.1
- Claude Fable 5
- Claude Opus 5.5
- Claude Opus 5
- Claude Opus 4.8
- Claude Sonnet 5.5
- Claude Haiku 5.5
- GPT-6 Astra
- GPT-6.1 Sol
- Gemini 4 argon
- Claude Sonnet 5.5 Slow
- GPT-6 Astra Slow
A check means the model is included in the plan. A dash means it is not available.
Ultra top-up PLN 20.00 = 200 Ultra credits
Needs an active plan. Used only after your plan limit runs out and does not expire on the 14-day renewal.
No refunds. Prices may change without notice. Terms of service
Slow models
Limit gone? Keep working.
When you use up your monthly limit, you do not have to wait for next month. Switch to a model ending in -slow and keep working.
▐▛███▜▌ ▝▜█████▛▘ ▘▘ ▝▝
Claude Code v2.1.312
Opus 5.5 with medium effort · API Usage Billing
~\my-app
❯ add rate limiting to the /login route
● API Error: 429 monthly limit exhausted
⎿ switch to a -slow model to keep working
⏵⏵ auto mode on (shift+tab to cycle)
What to know
- Slow models are slower. They wait at the very end of the queue, behind every other request, so answers take longer.
- They work only after your monthly limit runs out. While you still have a limit, use the regular models.
- Every slow request takes a bit from your Ultra pool, at least one minimal unit. When Ultra runs out, slow stops working too.
- Context is 200k tokens, no 1M variant.
- You can run one slow request at a time.
- The Lite plan has no slow models.
- claude-sonnet-5.5-slowfrom plan Pro
- gpt-6-astra-slowfrom plan Max x3
Official API vs Epsilon
Same model. Different bill.
Claude Opus 5.5 [1m], 2 500 requests a month. Each one: 200k cached context read, 10k cache write, 3k output.
Official API (per token)
~PLN 1,474.00
$375 / month · $0.15 / request
- input $4 · output $20 / MTok
- cache write $5 · cache read $0.2 / MTok
Epsilon (per request)
PLN 80.00
Plan Pro, $6 000 limit / month
- This example uses only ~6% of the limit
- Cache and 1M context included
API prices from the official Anthropic price list, USD at ~3.93 PLN. Example only, real usage varies.
FAQ
Questions people ask.
Short answers about plans, requests and payments. Missing something? Write on Discord.
Ask on DiscordYes. Use the key in Claude Code (CLI, VS Code, desktop) or in EpsilonCLI, our own terminal agent. You get Claude, GPT and Gemini under one key and one monthly plan.
Plans start at 40 PLN for 30 days with a $1 000 monthly limit at API prices. Pro costs 80 PLN and gives $6 000. You pay in PLN, upfront, with no subscription.
Normal is your monthly limit in dollars: every request takes its API price from it, and it resets on the 1st at 02:00 Warsaw time. Ultra is a separate pool for harder requests that renews every 14 days. It is used when something goes wrong, for example Anthropic's Secure Guard blocks a request (Anthropic charges for it even before any output), another model error happens or someone tries to manipulate the model. You can buy 200 Ultra credits for 20 PLN.
From the Pro plan, once your monthly limit runs out you can switch to claude-sonnet-5.5-slow (from Max x3 also gpt-6-astra-slow) and keep working. They are slower, wait at the end of the queue, run one request at a time with 200k context and take a bit from Ultra on every request.
Yes. All prices are in PLN, plans are prepaid for 30 days and the Ultra top-up is a one-time purchase.
No. Plans are prepaid, nothing renews on its own and nothing charges your card again.
No. Digital content is delivered instantly, so all sales are final. Details in the Terms of service.
Log in with Discord, pick a plan, create a key and run epsilon settings. The whole thing takes about 5 minutes.