Snapshot rates may differ from current charges.
TRANSPARENT MODEL PRICING
Official rates and channel discounts across Claude Code, Codex, Gemini, Grok, and TypeSafe — including image and video generation prices.
Real multiplier = group multiplier ÷ average credits per $1
Channel price = official price × real multiplier
Discount = 1 − real multiplier, shown as a whole percentage rounded down
Example: Codex group multiplier 0.6 ÷ 7.09 = 0.0846, which is 91% off, so a $2 official price becomes $0.169.
Prices below use the selected average. Each average is taken across all top-up packages or all subscription plans; your actual rate depends on the package you buy.
Channel USD price is 90% off the official price (group multiplier 0.65) · billed per 1M tokens
Single channel tier · channel rates per 1M tokens
Text model prices are 90% off the official price. Nano Banana 2 image generation is billed per image below.
| Model | Input | Output | Cache write (5m) | Cache read | Discount |
|---|---|---|---|---|---|
| gemini-4-argonLow, medium and high use the same token rates; output includes thinking tokens. Official introductory rates through Nov 30, 2026. From Dec 1, 2026: $4.00 input / $20.00 output / $0.20 cache read per 1M tokens before the group multiplier. Implicit caching only; explicit caching is not supported. | $0.183/M$2.00 official | $0.917/M$10.00 official | —/M— official | $0.00917/M$0.10 official | |
| gemini-4-hydrogenLow, medium, high and xhigh (Extra) use the same token rates; output includes thinking tokens. Implicit caching only; explicit caching is not supported. | $0.138/M$1.50 official | $0.688/M$7.50 official | —/M— official | $0.00688/M$0.075 official | |
| gemini-3.8-flashOfficial introductory rates through Dec 31, 2026. From Jan 1, 2027: $1.50 input / $7.50 output / $0.15 cache read per 1M tokens before the group multiplier. Low, medium, high and tiered variants use the same token rates. | $0.0688/M$0.75 official | $0.344/M$3.75 official | —/M— official | $0.00688/M$0.075 official | |
| gemini-3.5-flash-lite | $0.0275/M$0.3 official | $0.229/M$2.5 official | —/M— official | $0.00275/M$0.03 official | |
| gemini-3.1-flash-lite | $0.0229/M$0.25 official | $0.138/M$1.5 official | —/M— official | $0.00229/M$0.025 official | |
| gemini-3.1-flash-lite-preview | $0.0229/M$0.25 official | $0.138/M$1.5 official | —/M— official | $0.00229/M$0.025 official | |
| gemini-3-flash | $0.0459/M$0.5 official | $0.275/M$3 official | —/M— official | $0.00459/M$0.05 official | |
| gemini-3-flash-preview | $0.0459/M$0.5 official | $0.275/M$3 official | —/M— official | $0.00459/M$0.05 official | |
| gemini-2.5-flash | $0.0275/M$0.3 official | $0.229/M$2.5 official | —/M— official | $0.00275/M$0.03 official | |
| gemini-2.5-flash-lite | $0.00917/M$0.1 official | $0.0367/M$0.4 official | —/M— official | $0.00092/M$0.01 official |
Nano Banana 2 · per-image pricing · Gemini API and AGY CLI
Supported models · per image
Image generation verified with gemini-3.1-flash-image (Nano Banana 2).
Platform billing is per generated image, 90% off the official price. Official references use $60 per 1M image output tokens: 1K = 1,120 tokens, 2K = 1,680 tokens, 4K = 2,520 tokens. Text/image input and text output within image generation requests are included without additional charges. Separate conversation requests use the token rates above. Use the Gemini API generateContent endpoint or AGY's built-in image-generator with the same Gemini API key and gateway.
| Nano Banana 2 output resolution | Official price | Platform price | Discount |
|---|---|---|---|
| 1K1,120 image output tokens · default | $0.0672 / image | $0.00616224/ image | |
| 2K1,680 image output tokens | $0.1008 / image | $0.00924336/ image | |
| 4K2,520 image output tokens | $0.1512 / image | $0.01386504/ image |
Snapshot captured . Model, context and channel rules apply; the console billing record determines actual usage.