Pricing breakdown

Every model is billed in credits— the same way for everyone. We take the model's real provider cost and convert it to credits with a fixed rate, so you only ever pay the true cost of what you run. Text models are billed on both the tokens you send (input) and the tokens the model writes back (output). No markup on models you don't touch.

At our standard rate, ৳100 ≈ 9,000 credits. The tables below show how many total tokens (input + output, assuming an even mix) that buys on each model.

Text & chat models

ModelContext windowCredits / 1k input tokensCredits / 1k output tokensTotal tokens per ৳100
Glm 5.3 Flash1.0M15~3,009,467
DeepSeek V4 Flash1M35~2,326,152
Qwen Auto1M26~2,445,154
Qwen Flash1M26~2,445,154
Llama 3 70B128K47~1,628,517
Qwen3.8 Flash1M39~1,575,872
Grok 4 Fast256K49~1,396,106
Llama 4 Scout320K49~1,396,106
Llama 4 Maverick1M411~1,221,747
Qwen3 Coder Next262K215~1,086,202
DeepSeek V366K417~888,757
Qwen3 Coder Flash1M418~814,369
Perplexity Sonar128K618~751,738
GPT 5.6 Luna922K422~698,134
DeepSeek V4128K622~651,596
Qwen Plus256K622~651,596
GPT 5.4 Nano272K423~674,082
Gemini 3.1 Flash Lite1.0M528~558,594
Grok Code Fast256K428~575,025
Qwen3.6 Plus1M635~444,285
GPT 5.4 Codex Mini400K537~434,428
Perplexity Sonar Reasoning128K937~390,986
DeepSeek R166K1040~356,731
Qwen Max256K1144~325,798
Gemini 2.5 Flash1.0M646~349,077
Gemini 3 Flash1M1146~315,297
Kimi K2256K1146~315,297
Kimi K2 Thinking262K1146~315,297
Kimi K2.5262K1155~271,497
Qwen3 Coder Plus1M1159~257,220
Gemini 3.6 Flash1.0M1469~217,204
Gemini 3.7 Flash1.0M1469~217,204
Gemini 3.8 Flash1.0M1469~217,204
Glm 5.31.0M2681~168,522
GPT 5.4 Mini272K1483~186,179
Claude Haiku 4.5200K1892~162,903
Grok 4.5500K37110~122,183
Grok 4.6500K37110~122,183
Qwen3.8 Max1M37110~122,183
Qwen3 Max Thinking262K22110~135,759
Qwen3.6 Max Preview262K18114~135,762
Gemini 3.5 Flash1.0M28166~93,089
Gemini 3.5 Flash Lite1.0M28166~93,089
GPT 4o128K46184~78,197
GPT 5.4 Codex400K23184~86,886
GPT 5.6 Sol922K37184~81,455
Claude Sonnet 51M37184~81,455
Gemini 3.1 Pro2M24184~86,501
GPT 5.6 Terra922K37221~69,819
GPT 5.5 Codex400K32258~62,061
GPT 5.41.1M46276~55,854
Claude Sonnet 4.61M55276~54,303
Grok 4256K55276~54,303
Kimi K31.0M55276~54,303
Claude Opus 4.71M92460~32,582
Claude Opus 4.81M92460~32,582
Claude Opus 51M92460~32,582
GPT 5.51.1M92552~27,928
Grok 4 Multi Agent256K110552~27,152
GPT 6 Astra922K184921~16,291
Claude Fable 51M184921~16,291
Claude Fable 5.11M184921~16,291

Image, video & audio models

Media models are billed per generation (per image, per second of video, or per clip) rather than per token.

ModelTypeCredits per generation
Flux SchnellImage55
Minimax SpeechAudio184
Qwen ImageImage460
MusicgenAudio552
Flux ProImage737
Recraft V3Image737
GPT Image 2Image737
Ideogram V3Image1,473
SeedanceVideo3,315
Wan VideoVideo4,604
Hailuo VideoVideo9,208
Sora 2Video27,623
Veo 3 FastVideo58,928
Sora 2 ProVideo82,868
Veo 3Video110,490

Credits come with every plan and package — pick the one that fits.

See pricing