Pricing

What	Cost
Chat	metered $/MTok (input + output; discounted rate for cached input) — see the rate table below
Images	$0.03 – $0.16/image
Video	$3.00 – $5.00/video
Search	$0.05/query
Estimates, status polling, model list	Free

Every response prices itself — no spreadsheet required:

{ "cost": { "this_request": "$0.0187" }, "balance": "$4.820" }

Chat is billed by your exact token usage, not a flat per-request fee: a request preauthorizes an estimated maximum against your balance and settles down to what you actually used once the response completes. There’s no length-based rejection — a long-context request just needs balance to cover it. See GET /api/v1/pricing for the live, machine-readable rate sheet — it’s generated from the same code that bills you, so it can never drift from this page.

Developer Balance

You prepay a balance and requests draw from it — the same balance covers the REST API and the MCP server, and it never expires. Add funds at Developers → Billing (packs start at $10), or verify your phone on your first paid action to unlock a $0.50 starter balance free. What a $10 pack buys: pricing scales with the model and your usage, not a flat tier — as a range, that’s roughly ~5,500 typical requests on our most efficient models (e.g. llama-4-scout, ~$0.0018/typical request) down to ~27 on our most capable (gpt-5.4-pro, ~$0.37/typical request). “Typical” = a 5K-input/1K-output request; your real mileage depends on your actual token counts.

Every model's exact price

Chat — `POST /api/v1/chat`

Billed per token: metered_input_per_mtok + metered_output_per_mtok from GET /api/v1/pricing are the real rates. The typical price below is a derived reference at 5K input / 1K output tokens — useful for comparing models, not what you’re actually charged (that scales with your real usage). Provider prompt-cache reads bill at a discounted rate; a few large-context models switch to a whole-request tier once input crosses a threshold — both are in the live rate sheet.

Model	Provider	Typical price
`gpt-5.4-pro`	OpenAI	~$0.3696
`claude-opus-4.6`	Anthropic	~$0.056
`gpt-5.4`	OpenAI	~$0.0308
`claude-sonnet-4.6`	Anthropic	~$0.0336
`claude-sonnet-4.5`	Anthropic	~$0.0336
`mistral-large`	Mistral AI	~$0.0192
`gemini-2.5-pro`	Google	~$0.0182
`gpt-5`	OpenAI	~$0.0255
`gemini-3-pro`	Google	~$0.0246
`gemini-3.1-pro`	Google	~$0.0246
`claude-haiku-4.5`	Anthropic	~$0.0113
`o3-mini`	OpenAI	~$0.0112
`grok-4`	xAI	~$0.0105
`glm-5`	Zhipu AI	~$0.0071
`gemini-3-flash`	Google	~$0.0066
`deepseek-v3`	DeepSeek	~$0.0064
`kimi-k2`	Moonshot AI	~$0.0064
`gemini-2.5-flash`	Google	~$0.0051
`kimi-k2.5`	Moonshot AI	~$0.0049
`gpt-5-mini`	OpenAI	~$0.0042
`uncensored-ai`	—	~$0.0045
`ninja-1`	ByteDance	~$0.0045
`llama-4-maverick`	Meta	~$0.0031
`minimax-m2.5`	MiniMax	~$0.0029
`qwq-32b`	Alibaba	~$0.0029
`llama-4-scout`	Meta	~$0.0018

Always read live numbers from GET /api/v1/pricing — this table is a snapshot and will drift as rates or models change.

Images — `POST /api/v1/images`

Price	Models
$0.16	`gpt-image-2`
$0.10	`flux-kontext-max` `nano-banana-pro`
$0.08	`flux-2-flex` `flux-1-pro-ultra` `recraft-v3` `google-imagen-4` `seedream`
$0.07	`flux-1-fill`
$0.06	`nano-banana-2`
$0.05	`flux-2-pro` `flux-kontext-pro`
$0.03	`flux-2-klein` `nano-banana`

n images = n × the per-image price.

Video — `POST /api/v1/video`

Price	Models
$5.00	`veo-3.1` `google-veo-2`
$3.00	`seedance-2` `kling-video` `veo-3.1-fast` `google-veo-3-fast`

Failed video jobs are refunded automatically.

Search — `POST /api/v1/search`

$0.05 per query.

You’ll never be surprised

Price it first

/estimate is free and returns the estimated maximum cost — with cheaper alternatives.

Warnings before walls

Below $5.00 balance, every response carries a balance_warning; below $1.00 it becomes critical_low_balance. Out of funds = a clean 402 before anything runs.

Failures don't bill

Failed chat calls, failed batch items, and failed videos are not charged — video refunds are automatic.

Watch it live

Per-request logs and spend in the console.

From code

# Balance
r = requests.get("https://www.ninjachat.ai/api/api-keys", headers=HEADERS)
print(r.json()["billing"]["balanceFormatted"])   # "$42.50"

# Usage logs — filter by group/model/status/date, paginate with limit/offset
r = requests.get("https://www.ninjachat.ai/api/api-keys/logs",
    headers=HEADERS, params={"limit": 50, "group": "chat"})

# CSV export
# GET /api/api-keys/logs?format=csv&from=2026-03-01

# Purchase history
r = requests.get("https://www.ninjachat.ai/api/api-keys/transactions", headers=HEADERS)

Get Started

Chat

Superpowers

Media & Search

Reference

Developer Balance

Chat — `POST /api/v1/chat`

Images — `POST /api/v1/images`

Video — `POST /api/v1/video`

Search — `POST /api/v1/search`

You’ll never be surprised

Price it first

Warnings before walls

Failures don't bill

Watch it live

From code

​Developer Balance

​Chat — POST /api/v1/chat

​Images — POST /api/v1/images

​Video — POST /api/v1/video

​Search — POST /api/v1/search

​You’ll never be surprised

Price it first

Warnings before walls

Failures don't bill

Watch it live

​From code

Developer Balance

Chat — `POST /api/v1/chat`

Images — `POST /api/v1/images`

Video — `POST /api/v1/video`

Search — `POST /api/v1/search`

You’ll never be surprised

From code