How does NovAI compare to OpenRouter in pricing?

NovAI charges zero platform fee on top of vendor pricing, while OpenRouter adds a 5% markup. For Doubao, DeepSeek, Qwen, GLM, and Moonshot models, NovAI offers the lowest per-token cost with OpenAI-compatible API.

Why choose NovAI over direct vendor APIs?

NovAI provides a unified OpenAI-compatible endpoint for multiple Chinese AI vendors, Hong Kong-based low-latency routing, USDT (TRC20) billing, and free $2.00 credit for new users. No separate accounts per vendor required.

What are the 2026 AI API pricing trends?

2026 has seen Doubao-Seed-2.0 at $0.40/1M input, DeepSeek-V4-Pro at $0.57/1M, Qwen3.6-Max at $0.48/1M. Chinese models now match GPT-4-level quality at roughly 1/10th the cost, with 128K-256K context windows standard.

Simple, Pay-as-You-Go Pricing

Watermark notice — Every image model on NovAI is served by a Chinese provider (Zhipu AI, ByteDance, Tencent) and burns a visible AI-generated label such as AI生成 into the corner of the output under China’s labelling rules — NovAI cannot switch it off. Video output varies by model: in our 4 August 2026 tests Seedance 2.0 and CogVideoX-Flash showed no visible label, while Hunyuan Video 1.5 is labelled. Text and chat models are not affected. Which models, and what it rules out →

Five free-forever models included (text, image & video). Doubao exclusive. No monthly fees. Compare our prices vs OpenRouter below.

5 Free Models 38 Models Available OpenAI Compatible Hong Kong Servers

Beyond price — NovAI vs OpenRouter

Three dimensions that matter for production workloads with Chinese models.

Dimension	NovAI	OpenRouter
Chinese model day-one availability	Day-one launches (Qwen3.8-Max, GLM-5.2) + exclusives	Days-weeks delay; some models never listed
Media generation coverage	19 video & image models (Seedance, Hailuo, Seedream, Hunyuan, Cog) + free tier	Limited Chinese media models, no free tier
Payment methods	USDT (TRC20) — works without a credit card	Card-only

Model Pricing vs OpenRouter

All prices in USD per 1 million tokens. Green = cheaper than OpenRouter.

Model	Provider	NovAI Input	NovAI Output	OpenRouter Input	OpenRouter Output	Context	Notes
🎁 Free Tier
GGLM-4.6V-Flash	Zhipu AI	FREE	FREE	FREE	FREE	128K	Free Tier
🔥 ByteDance Doubao — NovAI Exclusive
BDoubao-Seed-2.0-Pro	ByteDance	$0.40	$2.00	N/A	N/A	128K	⭐ Flagship
BDoubao-Seed-2.0-Lite	ByteDance	$0.075	$0.45	N/A	N/A	128K	Fast & Cheap
BDoubao-Seed-2.0-Code	ByteDance	$0.448	$2.24	N/A	N/A	128K	Code Specialist
🔵 DeepSeek V4 Series
DDeepSeek-V4-Pro	DeepSeek	$0.57	$1.15	$0.57	$1.15	128K	Most Popular
DDeepSeek-V4-Flash	DeepSeek	$0.08	$0.17	$0.08	$0.17	128K	Best Value
DDeepSeek-V4.1-Flash	DeepSeek	$0.308	$1.231	N/A	N/A	128K	New Fast reasoning for chat & agents
DDeepSeek-V4-Flash-Vision	DeepSeek	$0.308	$1.231	N/A	N/A	128K	New Vision-language (experimental)
🟣 Alibaba Qwen Series
QQwen3.7-Max	Alibaba Cloud	$1.16	$3.48	~$1.48	~$4.43	128K	⭐ Newest
QQwen3.6-Max-Preview	Alibaba Cloud	$1.03	$6.16	N/A	N/A	128K	Preview
QQwen3-Max	Alibaba Cloud	$0.57	$2.65	N/A	N/A	32K	Stable
QQwen-Plus	Alibaba Cloud	$0.185	$0.53	~$0.26	~$0.78	128K	Budget
🟢 Zhipu GLM Series
GGLM-5.3-Flash	Zhipu AI	$0.08	$0.28	$0.075	$0.25	1M	New Budget flagship
GGLM-5.2	Zhipu AI	$1.19	$3.74	~$1.19	~$3.74	128K	New
GGLM-5V-Turbo	Zhipu AI	$1.08	$3.69	N/A	N/A	128K	Vision
GGLM-5.1	Zhipu AI	$1.05	$3.96	N/A	N/A	128K
GGLM-5	Zhipu AI	$0.60	$1.92	~$0.60	~$1.92	32K
GGLM-4.6V	Zhipu AI	$0.30	$0.90	N/A	N/A	128K	Vision
🟠 MiniMax
MMiniMax-Text-01	MiniMax	$0.14	$1.10	N/A	N/A	1M	1M Context
🔷 Moonshot Kimi
KMoonshot-v1-128k	Moonshot	$0.60	$2.50	N/A	N/A	128K	Document Analysis
KKimi-K2.8-Preview	Moonshot	$0.93	$3.89	N/A	N/A	256K	New Agentic preview
🔶 Tencent Hunyuan
HHunyuan-Hy4-Preview	Tencent	$0.834	$2.501	$0.834	$2.501	1M	New Flagship preview
HHunyuan-T1-Vision	Tencent	$0.55	$1.64	N/A	N/A	—	New Vision reasoning
HHunyuan-MT2-Plus	Tencent	$0.072	$0.287	N/A	N/A	8K	New Translation
HHunyuan-MT2-Lite	Tencent	$0.043	$0.172	N/A	N/A	8K	New Translation budget
HHY-Vision-2.0-Instruct	Tencent	$1.154	$2.692	N/A	N/A	—	New Fast multimodal vision (OCR, VQA)
HHunyuan-Turbos-Vision-Video	Tencent	$0.462	$1.385	N/A	N/A	—	New Image & video frame understanding
HHY-Role	Tencent	$0.369	$1.477	N/A	N/A	32K	New Role-play & character
🟧 Xiaomi MiMo
MMiMo-V2.5-Pro	Xiaomi	$0.427	$0.855	$0.435	$0.87	1M	New Reasoning

OpenRouter prices as of March 2026. Token counts based on UTF-8 encoding.

Embeddings Pricing per 1M input tokens

Text & multimodal vectors for RAG, semantic search and clustering. Billed on input tokens only.

Model	Vendor	Modality	Dims	NovAI Input
Tencent Kinfra Series
KKinfra-Text-Embedding-4B	Tencent	Text	2560	$0.0923 / 1M	Details
KKinfra-Text-Embedding-0.6B	Tencent	Text	1024	$0.0769 / 1M	Details
KKinfra-VL-Embedding-8B	Tencent	Multimodal	—	$0.0923 / 1M	Details
KKinfra-VL-Embedding-2B	Tencent	Multimodal	—	$0.0769 / 1M	Details

Multimodal models embed text, images and video frames into a shared vector space via POST /v1/embeddings.

Image Generation Pricing vs OpenRouter

Per-image billing, any resolution. Same models, lower prices. TokenHub models (Hunyuan / Vidu / Seedream / WAND Vega) are billed by real upstream token usage; the per-image figures are reference prices at base resolution.

Model	Vendor	NovAI	OpenRouter	Savings
ByteDance Seedream Series
SSeedream 4.5	ByteDance	$0.034 / image	$0.04 / image	-15%	Details
SSeedream 4.0	ByteDance	$0.027 / image	N/A	Exclusive	Details
TokenHub Image Models — Hunyuan / Vidu / Seedream / WAND Vega (token-billed, New)
HHunyuan Image 3.0	Tencent	~$0.031 / image	N/A	Exclusive	Details
VVidu Image Q2	Vidu	from ~$0.029 / image	N/A	Exclusive	Details
SSeedream v5.0 Pro (TokenHub)	ByteDance	from ~$0.046 / image	N/A	Exclusive	Details
SSeedream v5.0 Lite (TokenHub)	ByteDance	~$0.034 / image	N/A	Exclusive	Details
WWAND Vega Image Lite	Tencent	from ~$0.025 / image	N/A	Exclusive	Details
WWAND Vega Image Flash	Tencent	from ~$0.069 / image	N/A	Exclusive	Details
WWAND Vega Image Pro	Tencent	from ~$0.146 / image	N/A	Exclusive	Details

Generate images up to 4K resolution. Text-to-Image and Image-to-Image with reference images supported. TokenHub image models are billed by actual upstream total_tokens; per-image figures are base-resolution references that scale with resolution & reference count.

Video Generation Pricing vs OpenRouter

Seedance bills per second (resolution-based); TokenHub video models (MiniMax Hailuo, Hunyuan) are token-billed at $1.54 / 1M tokens — an estimate is pre-deducted at submit and reconciled to the real token count on success. Multi-modal input for maximum creative control.

Model	Vendor	Resolution	NovAI	OpenRouter	Savings
ByteDance Seedance Series
SSeedance 2.0	ByteDance	480p	$0.067 /s	$0.067 /s	Same price	Details
		720p	$0.163 /s	$0.199 /s	-18%
		1080p	$0.329 /s	$0.496 /s	-34%
		4K	$1.00 /s	$1.36 /s	-26%
SSeedance 2.0 Fast	ByteDance	480p	$0.054 /s	$0.054 /s	Same price	Details
		720p	$0.131 /s	$0.121 /s	Varies
		1080p	$0.264 /s	$0.302 /s	-13%
		4K	$0.802 /s	$0.625 /s	Varies
SSeedance 1.0 Pro			$2.05 / 1M tokens	N/A	Exclusive	Details
MiniMax Hailuo & Hunyuan Video — TokenHub (token-billed, New)
MHailuo H3	MiniMax	768P/1080P · T2V	$1.54 / 1M tok ≈$0.077 /s · $0.385 /5s	N/A	Exclusive	Details
MHailuo H3 Max	MiniMax	768P/1080P · T2V	$1.54 / 1M tok ≈$0.077 /s · $0.385 /5s	N/A	Exclusive	Details
MHailuo 2.3	MiniMax	768P/1080P · T2V	$1.54 / 1M tok ≈$0.062 /s · $0.308 /5s	N/A	Exclusive	Details
MHailuo 2.3 Fast	MiniMax	768P/1080P · I2V only	$1.54 / 1M tok ≈$0.042 /s · $0.208 /5s	N/A	Exclusive	Details
HHunyuan Video 1.5	Tencent	720p/1080p · T2V + I2V	$1.54 / 1M tok ≈$0.046 /s · $0.231 /5s	N/A	Exclusive	Details

4-15 second videos. Text, image, video, and audio reference inputs. Native audio sync on Seedance 2.0. Hailuo / Hunyuan run through the async POST /v1/video/generations task API (submit → poll → mp4 URL); their per-second figures are reference estimates derived from each model's token rate.

Audio Pricing vs OpenRouter

Token-billed at $1.54 / 1M tokens (10元/M upstream — the same rate as image & video). MiniMax Speech TTS & Music via POST /v1/audio/speech; Hunyuan / WAND speech-to-text via POST /v1/audio/transcriptions. Per-unit figures are reference prices that scale with text length, track length or audio duration.

Model	Vendor	Type	NovAI	OpenRouter	Savings
MiniMax Speech — Text-to-Speech (TTS) (token-billed, New)
MMiniMax Speech 2.8 HD	MiniMax	TTS · url/binary	$1.54 / 1M tok ≈$0.054 / 1K chars	N/A	Exclusive	Details
MMiniMax Speech 2.8 Turbo	MiniMax	TTS · url/binary	$1.54 / 1M tok ≈$0.031 / 1K chars	N/A	Exclusive	Details
MMiniMax Speech 2.6 HD	MiniMax	TTS · url/binary	$1.54 / 1M tok ≈$0.054 / 1K chars	N/A	Exclusive	Details
MMiniMax Speech 2.6 Turbo	MiniMax	TTS · url/binary	$1.54 / 1M tok ≈$0.031 / 1K chars	N/A	Exclusive	Details
MMiniMax Speech 02 HD	MiniMax	TTS · url/binary	$1.54 / 1M tok ≈$0.054 / 1K chars	N/A	Exclusive	Details
MMiniMax Speech 02 Turbo	MiniMax	TTS · url/binary	$1.54 / 1M tok ≈$0.031 / 1K chars	N/A	Exclusive	Details
MiniMax Music — Song Generation (token-billed, New)
MMiniMax Music v3.0	MiniMax	Music · url	$1.54 / 1M tok ≈$0.154 / song	N/A	Exclusive	Details
MMiniMax Music v2.6	MiniMax	Music · url	$1.54 / 1M tok ≈$0.154 / song	N/A	Exclusive	Details
Speech-to-Text (ASR) — Hunyuan / WAND (token-billed, New)
HHunyuan ASR 3.0	Tencent	ASR · text	$1.54 / 1M tok ≈$0.003 / min audio	N/A	Exclusive	Details
WWAND ASR v1	Tencent	ASR · text	$1.54 / 1M tok ≈$0.007 / min audio	N/A	Exclusive	Details

TTS returns an audio URL (response_format=url) or binary bytes (mp3/wav/opus/…); ASR returns text with timestamps & SRT subtitles. All audio is billed by actual upstream total_tokens × $1.54 / 1M; per-unit figures are references that scale with content length.

3D Generation Pricing vs OpenRouter

Flat per-generation price (default tier) — pre-deducted at submit, refunded in full if the job fails or is cancelled. Tencent Hunyuan 3D & Tripo (VAST) via the async POST /v1/three_d/generations task API (submit → poll → downloadable glb/obj/fbx/zip). TokenHub bills 3D upstream by credits and returns no token usage, so NovAI charges a flat price that always matches cost.

Model	Vendor	Input	NovAI	OpenRouter	Savings
Tencent Hunyuan 3D (flat per-generation, New)
HHunyuan 3D Express	Tencent	Text / Image	$0.277 / gen obj (zip)	N/A	Exclusive	Details
HHunyuan 3D 3.0	Tencent	Text / Image	$0.369 / gen obj (zip) + glb	N/A	Exclusive	Details
HHunyuan 3D 3.1	Tencent	Text / Image	$0.369 / gen obj (zip) + glb	N/A	Exclusive	Details
HHunyuan 3D Motion	Tencent	Text	$0.185 / gen fbx	N/A	Exclusive	Details
HHunyuan 3D PolyGen (Image)	Tencent	Image (required)	$0.554 / gen glb + obj + fbx	N/A	Exclusive	Details
Tripo (VAST) 3D (flat per-generation, New)
TTripo 3D 3.1	Tripo	Text / Image	$0.323 / gen glb	N/A	Exclusive	Details
TTripo 3D P1	Tripo	Text / Image	$0.323 / gen glb	N/A	Exclusive	Details

Submit returns a task id (status: queued); poll GET /v1/three_d/generations/{id} until succeeded to get model_url (primary, glb preferred) and outputs (every format). Generation typically takes 1–8 minutes. Failed/cancelled jobs are auto-refunded.

Why Choose NovAI?

Same models, better experience for international users.

NovAI

GLM-4.6V-FlashFREE

Doubao Seed 2.0 Pro$0.40 / $2.00

DeepSeek V4 Pro$0.57 / $1.15

PaymentUSDT (TRC20)

Sign upEmail only

Asia latencyHK servers

Free tier1 model

OpenRouter

GLM-4.6V-FlashFREE

GLM-5$0.60 / $1.92

PaymentCredit card

Sign upEmail + OAuth

Asia latencyUS servers

Free tierNone

GPT-4o (OpenAI)

Similar model$2.50 / $10.00

PaymentCredit card

Sign upPhone required

Asia latency~200ms TTFT

Free tierLimited

Use Case	NovAI	OpenRouter	OpenAI GPT-4o	Savings
1K chat messages/day (30 days)	$0.00 (free models)	$3.00	$37.50	100% vs OR
10K code reviews/month (DeepSeek)	$6.40	$6.40	$125.00	95% vs GPT
100 docs/day (GLM-5)	$2.98	$3.02	$45.00	1% vs OR
Long doc analysis (MiniMax 1M)	$1.80	$1.80	N/A (128K limit)	Best value

Pricing FAQ

How does billing work?

NovAI uses a prepaid balance system. Top up your account with USDT (TRC20) (minimum $5), and usage is deducted per token. You can check your balance and usage history in the dashboard at any time.

Is there a free tier?

Yes! GLM-4.6V-Flash is completely free with no usage limits. No credit card required. Start using it immediately after signup.

How do your prices compare to OpenRouter?

Our prices are competitive with OpenRouter, and we offer Doubao Seed 2.0 series exclusively — not available on OpenRouter. GLM-5 models are cheaper. Plus, we offer USDT (TRC20) payments and Hong Kong servers for better Asia-Pacific latency.

Do I need a Chinese phone number?

No. NovAI handles all upstream authentication. You sign up with just an email address and get instant API access. No phone verification, no ID check, no credit card required.

Can I switch models anytime?

Yes. All models share the same API key and endpoint. Just change the "model" parameter in your API call. You can route different tasks to different models for optimal cost-performance balance.