deepseek-v4-flash
Standard pricingInput cache hitRp 120
Input cache missRp 2,100
OutputRp 8,500
AshProx
Simple, transparent pricing
Use the models you need, at rates lower than you can imagine.
Pay only for the tokens you use. Prices below are base rates per 1,000,000 tokens.
Current rates
Model pricing
Pay-as-you-go rates per 1,000,000 tokens. Switch currency to see model prices converted at the planning rates (1 MYR = IDR 4,000 · 1 THB = IDR 500 · 1 VND = IDR 0.7 · 1 INR = IDR 200).
deepseek-v4-flash-vision-exp
Standard pricingInput cache hitRp 150
Input cache missRp 2,500
OutputRp 8,600
deepseek-v4-pro
Standard pricingInput cache hitRp 220
Input cache missRp 3,500
OutputRp 8,750
gemini-3.7-flash
Standard pricingInput cache hitRp 1,275
Input cache missRp 12,750
OutputRp 63,750
gemini-3.1-pro
Standard pricingInput cache hitRp 1,275
Input cache missRp 12,750
OutputRp 63,750
gpt-5.5
Context-aware pricingRequests above 200,000 context tokens use long-context rates.
short context
Input cache hitRp 1,806
Input cache missRp 18,063
OutputRp 108,375
long context
Input cache hitRp 3,613
Input cache missRp 36,125
OutputRp 162,563
gpt-5.4
Context-aware pricingRequests above 200,000 context tokens use long-context rates.
short context
Input cache hitRp 903
Input cache missRp 9,031
OutputRp 54,188
long context
Input cache hitRp 1,806
Input cache missRp 18,063
OutputRp 81,281
gpt-5.4-mini
Standard pricingInput cache hitRp 271
Input cache missRp 2,710
OutputRp 16,256
gpt-5.6-sol
Context-aware pricingRequests above 200,000 context tokens use long-context rates.
short context
Input cache hitRp 3,400
Input cache missRp 42,500
OutputRp 204,000
long context
Input cache hitRp 6,800
Input cache missRp 85,000
OutputRp 306,000
gpt-5.6-terra
Context-aware pricingRequests above 200,000 context tokens use long-context rates.
short context
Input cache hitRp 1,360
Input cache missRp 17,000
OutputRp 81,600
long context
Input cache hitRp 2,720
Input cache missRp 34,000
OutputRp 122,400
gpt-5.6-luna
Context-aware pricingRequests above 200,000 context tokens use long-context rates.
short context
Input cache hitRp 136
Input cache missRp 1,700
OutputRp 8,160
long context
Input cache hitRp 272
Input cache missRp 3,400
OutputRp 12,240
Monthly plans
Regional pricingSubscription access
Subscriptions include a monthly weighted-token allowance. Expensive models and fast requests consume more allowance. Plans are assigned manually and stop when the allowance is exhausted.
| Plan | Allowance | Limit | Indonesia IDR | Malaysia MYR | Thailand THB | Vietnam VND | India INR |
|---|---|---|---|---|---|---|---|
| Starter | 50M tokens | 30 req/min | IDR 99,000/ month | MYR 29/ month | THB 199/ month | VND 149,000/ month | INR 799/ month |
| Standard | 100M tokens | 60 req/min | IDR 224,100/ month | MYR 62/ month | THB 449/ month | VND 359,100/ month | INR 1,299/ month |
| Pro | 250M tokens | 120 req/min | IDR 325,000/ month | MYR 81/ month | THB 650/ month | VND 464,286/ month | INR 1,625/ month |
Plan estimates
Not a guaranteeHow far could it go?
These are illustrative request estimates, not ceilings. Adjust uncached input from 100 to 200K tokens, output from 50 to 32K tokens, and cached input from 0 to 1M tokens to model different sessions. Actual usage depends on the model and conversation history.
| Model | Tier | Starter estimated requests | Standard estimated requests | Pro estimated requests |
|---|---|---|---|---|
| deepseek-v4-flash | Tiny | -- | -- | -- |
| deepseek-v4-flash-vision-exp | Tiny | -- | -- | -- |
| deepseek-v4-pro | Middle | -- | -- | -- |
| gemini-3.7-flash | Medium | -- | -- | -- |
| gemini-3.1-pro | Flagship | -- | -- | -- |
| gpt-5.5 | Middle | -- | -- | -- |
| gpt-5.4 | Tiny | -- | -- | -- |
| gpt-5.4-mini | Tiny | -- | -- | -- |
| gpt-5.6-sol | Flagship | -- | -- | -- |
| gpt-5.6-terra | Middle | -- | -- | -- |
| gpt-5.6-luna | Tiny | -- | -- | -- |
Tiny models stretch furthest, middle-tier models use the baseline, and GPT-5.6 Sol is weighted at 3× because it sits above the rest of the lineup. Estimates use the same weighted-token concept used for subscription limits.
Which is cheaper?
IDRSubscription vs pay-as-you-go
For a typical request (2,350 fresh + 327,000 cached input + 1,530 output tokens), the effective per-1M-weighted-token rate decides it. The cheapest plan is Pro at Rp 1,300 per 1M weighted tokens — a model above that rate is cheaper on a subscription, anything below is cheaper to pay per request.
| Model | Tier | Pay-as-you-go per 1M weighted tokens | Starter Rp 1,980 /1M | Standard Rp 2,241 /1M | Pro Rp 1,300 /1M | Cheapest option |
|---|---|---|---|---|---|---|
| deepseek-v4-flash | Tiny | Rp 346 | PAYG cheaper | PAYG cheaper | PAYG cheaper | Pay-as-you-go |
| deepseek-v4-flash-vision-exp | Tiny | Rp 412 | PAYG cheaper | PAYG cheaper | PAYG cheaper | Pay-as-you-go |
| deepseek-v4-pro | Middle | Rp 283 | PAYG cheaper | PAYG cheaper | PAYG cheaper | Pay-as-you-go |
| gemini-3.7-flash | Medium | Rp 3,291 | Plan cheaper | Plan cheaper | Plan cheaper | Subscription |
| gemini-3.1-pro | Flagship | Rp 548 | PAYG cheaper | PAYG cheaper | PAYG cheaper | Pay-as-you-go |
| gpt-5.5 | Middle | Rp 2,414 | Plan cheaper | Plan cheaper | Plan cheaper | Subscription |
| gpt-5.4 | Tiny | Rp 2,414 | Plan cheaper | Plan cheaper | Plan cheaper | Subscription |
| gpt-5.4-mini | Tiny | Rp 724 | PAYG cheaper | PAYG cheaper | PAYG cheaper | Pay-as-you-go |
| gpt-5.6-sol | Flagship | Rp 1,535 | PAYG cheaper | PAYG cheaper | Plan cheaper | Subscription |
| gpt-5.6-terra | Middle | Rp 1,842 | PAYG cheaper | PAYG cheaper | Plan cheaper | Subscription |
| gpt-5.6-luna | Tiny | Rp 368 | PAYG cheaper | PAYG cheaper | PAYG cheaper | Pay-as-you-go |
A subscription can still win if you're a heavy user of the models marked Subscription and your sessions re-read very little cached context — every fresh prompt costs more than the cached-average shown here. These are estimates for planning, not a guarantee.
How billing works
Input tokens are split between cache hits and cache misses. Cached input is cheaper. Output tokens have their own rate, and fast or priority service uses the multiplier shown above.