Skip to content
APMIX.AI
EN

Pricing

Simple plans. Monthly or yearly.

Pick a monthly allowance of weighted tokens and spend it on any model in your plan, from one key. Each model has its own multiplier; the table further down turns your allowance into the real tokens it buys on every one of them.

Billing period

Billed monthly. Cancel any time.

Prices in USD. Taxes may apply depending on your country. Subscribers can add one-time usage top-ups from $1.49 when a month runs short.

How it works

One allowance. One multiplier per model.

Your plan is a monthly budget of weighted tokens. Every model converts the tokens you actually send and receive into weighted tokens at its own rate, so a model with a lower multiplier stretches your allowance further.

  1. 01

    Choose an allowance

    Starter, Pro or Max: 60M, 200M or 500M weighted tokens a month, shared across every model in the plan and every key on the account.

  2. 02

    Every model has a multiplier

    Each model's multiplier follows what it costs to run: under 1× on many models, several times more on others. Prompt and completion tokens, reasoning included, are counted and multiplied by that rate.

  3. 03

    Divide to get real tokens

    Allowance ÷ multiplier is what you can actually send to that model in a month. At 1× that is the allowance itself; at 0.5×, twice as much; at 4×, a quarter of it.

Max allowance: 500M weighted tokens a month ÷ Model multiplier 1.3× = 384.6M Real tokens a month on that model.

Model rates

51 models. One multiplier each.

The multiplier is how many weighted tokens one real token costs on a model. The plan columns show the real tokens a full monthly allowance buys there — the number you can actually send.

Plan
Provider
Showing 51 of 51

Included from Starter20 models

  • Grok 4.5

    grok-4.5xAI

    0.5×1K tokens → 500 weightedContext 500K
    Starter
    120M
    Pro
    400M
    Max
    1B
  • DeepSeek V4 Pro

    deepseek-v4-proDeepSeek

    0.5×1K tokens → 500 weightedContext 1M
    Starter
    120M
    Pro
    400M
    Max
    1B
  • DeepSeek V4 Flash

    deepseek-v4-flashDeepSeek

    0.7×1K tokens → 700 weightedContext 1M
    Starter
    85.7M
    Pro
    285.7M
    Max
    714.3M
  • GLM-5.3 Flash

    glm-5.3-flashZ.ai (GLM)

    0.3×1K tokens → 300 weightedContext 1M
    Starter
    200M
    Pro
    666.7M
    Max
    1.7B
  • GLM-5.2

    glm-5.2Z.ai (GLM)

    1.5×1K tokens → 1.5K weightedContext 1M
    Starter
    40M
    Pro
    133.3M
    Max
    333.3M
  • Claude Sonnet 4.6

    claude-sonnet-4-6Anthropic

    1×1K tokens → 1K weightedContext 200K
    Starter
    60M
    Pro
    200M
    Max
    500M
  • DeepSeek V4.1 FlashNew

    deepseek-v4.1-flashDeepSeek

    0.7×1K tokens → 700 weightedContext 1M
    Starter
    85.7M
    Pro
    285.7M
    Max
    714.3M
  • Grok 4.6

    grok-4.6xAI

    0.5×1K tokens → 500 weightedContext 500K
    Starter
    120M
    Pro
    400M
    Max
    1B
  • MiniMax M3

    minimax-m3MiniMax

    0.3×1K tokens → 300 weightedContext 512K
    Starter
    200M
    Pro
    666.7M
    Max
    1.7B
  • GPT 5.6 Luna

    gpt-5.6-lunaOpenAI

    0.5×1K tokens → 500 weightedContext 1.1M
    Starter
    120M
    Pro
    400M
    Max
    1B
  • Grok 4.7New

    grok-4.7xAI

    0.5×1K tokens → 500 weightedContext 500K
    Starter
    120M
    Pro
    400M
    Max
    1B
  • GPT 6 LunaNew

    gpt-6-lunaOpenAI

    0.38×1K tokens → 380 weightedContext 1.1M
    Starter
    157.9M
    Pro
    526.3M
    Max
    1.3B
  • Muse Spark 1.3

    muse-spark-1.3Meta

    0.5×1K tokens → 500 weightedContext 1M
    Starter
    120M
    Pro
    400M
    Max
    1B
  • Claude Haiku 5.5New

    claude-haiku-5-5Anthropic

    0.9×1K tokens → 900 weightedContext 1M
    Starter
    66.7M
    Pro
    222.2M
    Max
    555.6M
  • Grok Build 0.1

    grok-build-0.1xAI

    0.5×1K tokens → 500 weightedContext 256K
    Starter
    120M
    Pro
    400M
    Max
    1B
  • Grok 4.3

    grok-4.3xAI

    0.5×1K tokens → 500 weightedContext 1M
    Starter
    120M
    Pro
    400M
    Max
    1B
  • MiniMax M3.1 FlashNew

    minimax-m3.1-flashMiniMax

    0.3×1K tokens → 300 weightedContext 1M
    Starter
    200M
    Pro
    666.7M
    Max
    1.7B
  • MiMo V2.6 FlashNew

    mimo-v2.6-flashXiaomi MiMo

    0.1×1K tokens → 100 weightedContext 1M
    Starter
    600M
    Pro
    2B
    Max
    5B
  • MiMo V2.6 ProNew

    mimo-v2.6-proXiaomi MiMo

    0.3×1K tokens → 300 weightedContext 1M
    Starter
    200M
    Pro
    666.7M
    Max
    1.7B
  • GPT 6.1 TerraNew

    gpt-6.1-terraOpenAI

    0.7×1K tokens → 700 weightedContext 1.1M
    Starter
    85.7M
    Pro
    285.7M
    Max
    714.3M

Included from Pro22 models

  • Gemini 3.8 Flash

    gemini-3.8-flashGoogle

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • Gemini 3.7 Flash

    gemini-3.7-flashGoogle

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • Gemini 3.6 Flash

    gemini-3.6-flashGoogle

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • Qwen3.8 Flash

    qwen3.8-flashAlibaba Qwen

    0.7×1K tokens → 700 weightedContext 1M
    Starter
    Not included
    Pro
    285.7M
    Max
    714.3M
  • Hunyuan 4 PreviewNew

    hy4-previewTencent Hunyuan

    1×1K tokens → 1K weightedContext 1M
    Starter
    Not included
    Pro
    200M
    Max
    500M
  • Claude Haiku 4.5

    claude-haiku-4-5Anthropic

    0.9×1K tokens → 900 weightedContext 200K
    Starter
    Not included
    Pro
    222.2M
    Max
    555.6M
  • GPT-5.6 Terra

    gpt-5.6-terraOpenAI

    1.3×1K tokens → 1.3K weightedContext 1.1M
    Starter
    Not included
    Pro
    153.8M
    Max
    384.6M
  • Claude Sonnet 5

    claude-sonnet-5Anthropic

    3×1K tokens → 3K weightedContext 1M
    Starter
    Not included
    Pro
    66.7M
    Max
    166.7M
  • Gemini 3.1 Pro

    gemini-3.1-proGoogle

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • GLM-5.3

    glm-5.3Z.ai (GLM)

    1.5×1K tokens → 1.5K weightedContext 1M
    Starter
    Not included
    Pro
    133.3M
    Max
    333.3M
  • GLM-5 Turbo

    glm-5-turboZ.ai (GLM)

    1×1K tokens → 1K weightedContext 202.8K
    Starter
    Not included
    Pro
    200M
    Max
    500M
  • Kimi K2.7 Code

    kimi-k2.7-codeMoonshot Kimi

    0.9×1K tokens → 900 weightedContext 262.1K
    Starter
    Not included
    Pro
    222.2M
    Max
    555.6M
  • Claude Opus 4.8

    claude-opus-4-8Anthropic

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • Claude Opus 4.7

    claude-opus-4-7Anthropic

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • Claude Opus 4.6

    claude-opus-4-6Anthropic

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M
  • GPT 6 SolNew

    gpt-6-solOpenAI

    3×1K tokens → 3K weightedContext 1.1M
    Starter
    Not included
    Pro
    66.7M
    Max
    166.7M
  • GPT 6.1 SolNew

    gpt-6.1-solOpenAI

    3×1K tokens → 3K weightedContext 1.1M
    Starter
    Not included
    Pro
    66.7M
    Max
    166.7M
  • Claude Sonnet 5.5New

    claude-sonnet-5-5Anthropic

    3×1K tokens → 3K weightedContext 1M
    Starter
    Not included
    Pro
    66.7M
    Max
    166.7M
  • Gemini 4 ArgonNew

    gemini-4-argonGoogle

    3.5×1K tokens → 3.5K weightedContext 1M
    Starter
    Not included
    Pro
    57.1M
    Max
    142.9M
  • Kimi K2.8 Preview

    kimi-k2.8-previewMoonshot Kimi

    0.9×1K tokens → 900 weightedContext 1M
    Starter
    Not included
    Pro
    222.2M
    Max
    555.6M
  • Muse WonderNew

    muse-wonder-1.0Meta

    3×1K tokens → 3K weightedContext 1M
    Starter
    Not included
    Pro
    66.7M
    Max
    166.7M
  • Kimi K3.1 ProNew

    kimi-k3.1-proMoonshot Kimi

    2×1K tokens → 2K weightedContext 1M
    Starter
    Not included
    Pro
    100M
    Max
    250M

Included from Max9 models

  • GPT-5.6 Sol

    gpt-5.6-solOpenAI

    4×1K tokens → 4K weightedContext 1.1M
    Starter
    Not included
    Pro
    Not included
    Max
    125M
  • GPT-5.5

    gpt-5.5OpenAI

    4×1K tokens → 4K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    125M
  • Qwen3.8 Max

    qwen3.8-maxAlibaba Qwen

    2.5×1K tokens → 2.5K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    200M
  • Claude Opus 5

    claude-opus-5Anthropic

    4×1K tokens → 4K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    125M
  • GPT-6 AstraNew

    gpt-6-astraOpenAI

    10×1K tokens → 10K weightedContext 1.1M
    Starter
    Not included
    Pro
    Not included
    Max
    50M
  • Claude Fable 5

    claude-fable-5Anthropic

    12×1K tokens → 12K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    41.7M
  • Kimi K3

    kimi-k3Moonshot Kimi

    2.5×1K tokens → 2.5K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    200M
  • Claude Fable 5.1

    claude-fable-5-1Anthropic

    12×1K tokens → 12K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    41.7M
  • Claude Opus 5.5New

    claude-opus-5-5Anthropic

    6×1K tokens → 6K weightedContext 1M
    Starter
    Not included
    Pro
    Not included
    Max
    83.3M

Multipliers are set per model and can change when a provider changes its prices. This page, each model's page and your dashboard always show the current value; requests already made keep the rate they were billed at.

apmix BusinessReselling AI, or building on it at volume?Business accounts get their own price per token, higher request limits, resale rights and a prepaid balance that never expires.Discover apmix Business

Pricing questions, answered.

Does every plan include every model?

Plans are tiered. Each plan includes the models in its own tier plus everything in the tiers below, so Max covers the whole catalog. The table above shows exactly which plans include a model, and the model's own page says the same. Every plan uses the same key and the same endpoint.

What are weighted tokens?

One unit across every model. Every prompt and completion token — reasoning tokens included — is multiplied by the model's rate and deducted from your monthly allowance. A 1,000-token exchange costs 1,000 weighted tokens on a 1× model, 500 on a 0.5× model and 4,000 on a 4× model.

Do yearly plans have a bigger allowance?

No. The allowance is monthly on both. A yearly plan is the same monthly allowance for twelve months at a lower price, paid once.

What happens when I use up the allowance?

Requests return a clear error until the monthly reset, or until you move to a bigger plan. An upgrade applies right away: until your next renewal, both the price difference and the extra tokens are pro rata, and from then on you get the bigger plan's full allowance. You can also set your own daily or weekly limits so nothing surprises you.

Can I buy more tokens without changing plan?

Yes. Subscribers can buy a usage top-up from Billing: 10M, 30M, 60M or 100M weighted tokens as a one-time payment, from $1.49 depending on your plan. Top-ups are spent after your monthly allowance, never expire, and can be paid by card, Apple Pay or Google Pay.

Can I change or cancel my plan?

Any time, from the dashboard. Upgrades take effect immediately. Downgrades and cancellations take effect at the end of the paid period, and your keys keep working until then.

Are there refunds?

Monthly plans are refundable in full as long as no more than 2.5M tokens have been used that month, on every plan. Yearly plans are not refundable. The refund policy has the details.

Start with any plan. Change it whenever you like.

Choose your allowance, create a key, and point your tools at apmix.

Get started