Progga
Start free

Price list

মূল্য তালিকা

Every model, every price, in taka.

Every price below is in BDT per 1 million tokens, about 750,000 words. Your prepaid balance works across both the API and Progga Chat, with no subscription.

Start with ৳50 credit
Input
Tokens you send, including instructions, conversation history, and other request content.
Output
Tokens generated by the model in its response.
Cached input
Repeated input reused from a provider cache and billed at a lower rate.

Available models

BDT per 1 million tokens. Prices can include up to six decimal places. Only active, currently routable models are listed.

  • DeepSeek V4.1 Flash

    deepseek-v4.1-flash
    Input
    ৳14.00
    Output
    ৳70.00
    Cached input
    ৳0.42

    178.3 tok/s TPS2.7 s TTFT

    Cache hit rate 25.23%

  • Deepseek v4 Flash 0731

    deepseek-v4-flash-0731
    Input
    ৳14.45
    Output
    ৳28.875
    Cached input
    ৳0.50

    149.7 tok/s TPS2.3 s TTFT

    Cache hit rate 87.94%

  • DeepSeek V4 Pro

    deepseek-v4-pro-0813
    Input
    ৳50.00
    Output
    ৳112.00
    Cached input
    ৳1.50

    Not enough recent data

  • glm-5.3

    glm-5.3
    Input
    ৳55.552
    Output
    ৳194.432
    Cached input
    ৳8.3328

    69 tok/s TPS1.7 s TTFT

    Cache hit rate 0%

  • glm-5.3-flash

    glm-5.3-flash
    Input
    ৳10.50
    Output
    ৳35.00
    Cached input
    ৳2.10

    34.4 tok/s TPS2.4 s TTFT

    Cache hit rate 0%

  • GLM-5.3-flash (fast)

    glm-5.3-flashx
    Input
    ৳14.00
    Output
    ৳49.00
    Cached input
    ৳2.80

    Not enough recent data

  • Kimi K2.7 Code

    kimi-k2.7-code
    Input
    ৳24.7375
    Output
    ৳104.1625
    Cached input
    ৳5.00

    89.1 tok/s TPS5.9 s TTFT

    Cache hit rate 89.64%

  • Kimi-K3

    kimi-k3
    Input
    ৳210.00
    Output
    ৳1,050.00
    Cached input
    ৳21.00

    Not enough recent data

  • Mimo v2.6 Flash

    −50%
    mimo-v2.6-flash
    Input
    ৳9.80৳19.60
    Output
    ৳19.60৳39.20
    Cached input
    ৳0.196৳0.392

    36.8 tok/s TPS8.2 s TTFT

    Cache hit rate 67.03%

  • muse-spark-1.3

    muse-spark-1.3
    Input
    ৳175.00
    Output
    ৳595.00
    Cached input
    ৳21.00

    Not enough recent data

  • muse-spark-1.3-contributor

    ⚡ Fastest
    muse-spark-1.3-contributor
    Input
    ৳14.00
    Output
    ৳28.00
    Cached input
    ৳0.50

    554.5 tok/s TPS3.8 s TTFT

    Cache hit rate 96.05%

  • Qwen3.8-Omni-Flash

    qwen3.8-omni-flash
    Input
    ৳21.00
    Output
    ৳65.80
    Cached input
    ৳2.24

    66.5 tok/s TPS3 s TTFT

    Cache hit rate 31.43%

Notes

How to read this

Metrics last calculated Nov 2, 2026, 12:00 PM.

Observed performance

  • Higher TPS (tokens per second) is faster. Lower TTFT (time to first token) is faster.
  • TPS and TTFT are observed medians over the last 24 hours of successful streaming API requests, refreshed every six hours.
  • Client network latency is excluded, so your own request latency will be higher and actual performance varies.
  • Not enough recent data means fewer than ten recent requests were available for that model.

Cache hit rate

Cache hit rate is the share of your input tokens the provider already had cached over the last 24 hours. Cached tokens are billed at the cached-input price above rather than the full input price, so a higher rate lowers your effective cost. How high it gets depends on your prompts: long, repeated prefixes — a fixed system prompt or an ongoing conversation — cache well, while short one-off requests may be below the provider's minimum cacheable size.

How billing works

Requests reserve enough balance before they start, then settle against the model's actual reported token usage. Unused reserved balance is released. Your Wallet shows the resulting BDT cost for every request.