SimianX AI LogoSimianX
Pricing Docs

Points & model rates

Every SimianX plan includes a points allowance. Points are only spent when AI actually computes for you — this page shows exactly how usage converts to points for every model we offer, and lets you estimate your own workload.

01

How billing works

Pick your models

Choose which AI model powers each agent — from lightweight, low-cost models to frontier flagships. Rates differ per model, never per feature.

Tokens are metered

When an agent runs, the model's real input and output tokens are counted — every call is visible in your usage history.

Tokens convert to points

Token usage converts straight into points at each model's fixed rate. No per-call fees, no minimum charge, no rounding up.

Browsing is free

Charts, candles, indicator data, leaderboards, and stories never consume points. Only AI computation does.

02

The formula

points = input tokens ÷ 1M × input rate + output tokens ÷ 1M × output rate

Input and output rates are each model's points per one million tokens — exactly the numbers in the rate tables below.

Worked example

One GPT-4o Mini call using 3,000 input + 1,000 output tokens: 3,000 ÷ 1M × 37.5 + 1,000 ÷ 1M × 150 ≈ 0.26 points.

03

Model rates

Points per one million tokens, by provider. This catalog is live — it updates when models are added or rates are adjusted.

Estimated points for one agent completing one analysis (~3.5K input + 1K output tokens — 30-day production average).

OpenAI

ModelInput · pts/1MOutput · pts/1M≈ pts / analysis
GPT-5.6 Sol1,2507,50011.9
GPT-5.6 Terra5003,0004.8
GPT-5.6 Luna503000.48
GPT-5.51,2507,50011.9
GPT-5.46253,7505.9
GPT-5.4 Mini187.51,1251.8
GPT-5.4 Nano50312.50.49
GPT-5.2437.53,5005.0
GPT-5.1312.52,5003.6
GPT-4o MiniDefault37.51500.28
GPT-4o6252,5004.7
GPT-4.1 Nano251000.19
GPT-4.1 Mini1004000.75
GPT-4.15002,0003.8

Anthropic

ModelInput · pts/1MOutput · pts/1M≈ pts / analysis
Claude Sonnet 4.57503,7506.4
Claude Haiku 4.52501,2502.1
Claude Opus 4.51,2506,25010.6
Claude Sonnet 4.67503,7506.4
Claude Opus 4.61,2506,25010.6
Claude Opus 4.71,2506,25010.6
Claude Opus 4.81,2506,25010.6
Claude Fable 52,50012,50021.3

Google

ModelInput · pts/1MOutput · pts/1M≈ pts / analysis
Gemini 3 Flash PreviewDefault2507501.6
Gemini 3.5 Flash3752,2503.6
Gemini 2.5 Flash756250.89
Gemini 2.5 Flash-Lite251000.19

DeepSeek

ModelInput · pts/1MOutput · pts/1M≈ pts / analysis
DeepSeek V4 Flash551650.36
DeepSeek V4 Pro1654951.1

xAI

ModelInput · pts/1MOutput · pts/1M≈ pts / analysis
Grok 4.65001,5003.3
Grok 4.55001,5003.3
Grok 4.3312.56251.7
Grok 4.20 Reasoning5001,5003.3
Grok 4.20 Non-Reasoning5001,5003.3
Grok 4.1 Fast Reasoning501250.30
Grok 4.1 Fast Non-Reasoning501250.30
Grok 4 Fast Reasoning501250.30
Grok 4 Fast Non-Reasoning501250.30
Grok 3 Mini751250.39
Grok 37503,7506.4

Qwen

ModelInput · pts/1MOutput · pts/1M≈ pts / analysis
Qwen Max5001,5003.3
Qwen 3 14B1253750.81
Qwen 3 8B1253750.81
04

Points calculator

Estimate what your workload costs. Presets use measured 30-day production averages; actual token usage varies with market conditions.

~3.5K input · 1K output tokens

Estimated total
2.81 points

0.28 points per run

What that means per plan
Free≈ 88 runs / day
Plus≈ 10,666 runs / month
Pro≈ 113,777 runs / month
05

Plan allowances

Free
25
points / day

+10 with Discord connected · +10 with a verified share on X

Plus
3,000
points / month
Pro
32,000
points / month
Compare plans
06

Good to know

Market data never costs points — charts, candles, indicators, leaderboards, and stories are always free to browse.

Autopilots and digests stop automatically when your points run out; at most the in-flight call completes.

Stronger models consume more points — a common setup is lightweight models for routine monitoring and a flagship model for final decisions.

When models are added or rates are adjusted, this page is refreshed. Points allowances reset daily on Free and with each billing cycle on Plus and Pro.

Rates last verified August 28, 2026