China's Frontier Models, One Marketplace

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 →

Video | Image | 3D | Chat | Embeddings | Audio — one API key, 105 models from 8 Chinese labs. Official-direct pricing with OpenRouter comparison built-in.

🔌 Official direct endpoints ⚡ Day-one model launches 🧾 Auditable billing 💳 USDT (TRC20)
105
Models
$0
Platform Fee
80ms
Avg Latency
$2
Free Credit
Browse by category
🎬 Video Generation API
Seedance, Hailuo H3, Hunyuan — from $0.042/s, free CogVideoX
🖼 Image Generation API
Seedream, Hunyuan, Vega — from $0.025/img, free CogView
🧊 Text-to-3D API
Hunyuan 3D & Tripo — glb/obj/fbx from $0.185
🔊 Speech & Audio API
MiniMax TTS & Music, Hunyuan/WAND ASR

Grouped by provider; flagship and higher-capability variants first. Product positioning, not an independent benchmark. Specialist tasks are not directly comparable; provider names refer to developers or published model brands, not API gateways.

ModelVendorContextNovAI / 1M tokensOpenRouter / 1M tokens
Alibaba14 models
Qwen3.8-MaxFlagship
Alibaba's newest flagship. Latest Qwen generation, strong CN & EN.
Alibaba1M$1.84 / $5.52$2.00 / $6.00
8% cheaper
Try
Qwen3.7-Max
Alibaba flagship. 128K ctx, strong CN & EN.
Alibaba128K$1.16 / $3.48$1.48 / $4.43
22% cheaper
Try
Qwen3.6-Max-Preview
Next-gen preview. Improved multilingual & reasoning.
Alibaba128K$1.03 / $6.16$1.03 / $6.16
≈ match
Try
Qwen3-Max
Solid all-rounder. Balanced performance & cost.
Alibaba128K$0.57 / $2.65$0.78 / $3.90
27% cheaper
Try
Qwen3 Coder Plus
Flagship agentic coding model. Repo-level reasoning, tool calling, 256K context.
Alibaba256K$0.65 / $3.25OpenRouter N/A
Official direct
Try
Qwen VL Max
Flagship vision-language model. Complex image reasoning, documents and charts.
Alibaba32K$0.80 / $3.20OpenRouter N/A
Official direct
Try
Qwen3 VL Plus
Vision-language model. Image understanding, OCR and visual question answering.
Alibaba256K$0.21 / $0.63OpenRouter N/A
Official direct
Try
Qwen3.5 OCR
Latest OCR model. Structured extraction, tables, handwriting and multilingual text.
Alibaba32K$0.70 / $0.70OpenRouter N/A
Official direct
Try
Qwen VL OCR
OCR-specialized vision model. High-accuracy text extraction from images and documents.
Alibaba32K$0.70 / $0.70OpenRouter N/A
Official direct
Try
Qwen MT Plus
Dedicated machine-translation model. 90+ languages, domain-aware terminology.
Alibaba8K$0.25 / $0.75OpenRouter N/A
Official direct
Try
Qwen-Plus
Budget-friendly. Great for high-volume simple tasks.
Alibaba128K$0.185 / $0.53$0.26 / $0.78
27% cheaper
Try
Qwen3 Coder Flash
Fast, low-cost coding model for high-volume developer workflows.
Alibaba256K$0.15 / $0.75OpenRouter N/A
Official direct
Try
Qwen3 VL Flash
Fast, cheap vision-language model for high-volume image analysis.
Alibaba256K$0.08 / $0.24OpenRouter N/A
Official direct
Try
Qwen3 Omni Flash
Omni-modal model. Text, image, audio and video input through one API.
Alibaba32K$0.15 / $0.60OpenRouter N/A
Official direct
Try
ByteDance5 models
Doubao Seed 2.1 ProFlagship
ByteDance flagship. 256K context, strong reasoning and agentic tool use.
ByteDance256K$0.40 / $2.00OpenRouter N/A
Official direct
Try
Doubao-Seed-2.0-Pro
3.2T MoE flagship. 256K ctx, GPT-4 class reasoning.
ByteDance256K$0.40 / $2.00OpenRouter N/ATry
Doubao-Seed-2.0-Code
Code generation specialist. Optimized for dev workflows.
ByteDance128K$0.448 / $2.24OpenRouter N/ATry
Doubao Seed 2.1 Turbo
Fast, cost-efficient Doubao variant for high-throughput production.
ByteDance256K$0.15 / $0.90OpenRouter N/A
Official direct
Try
Doubao-Seed-2.0-Lite
Cost-efficient low-latency workhorse. Strong multimodal.
ByteDance256K$0.075 / $0.45$0.25 / $2.00
70% cheaper
Try
DeepSeek4 models
DeepSeek V4.1 FlashFlagship
Fast, cost-efficient reasoning LLM for chat, code & agents.
DeepSeek128K$0.308 / $1.231OpenRouter N/A
Exclusive
Try
DeepSeek-V4-Pro
685B MoE, 128K ctx. Top-tier reasoning & coding.
DeepSeek128K$0.57 / $1.15$0.57 / $1.15
≈ match
Try
DeepSeek-V4-Flash
Fast distilled. 128K ctx, 3x cheaper than Pro.
DeepSeek128K$0.08 / $0.17$0.08 / $0.17
≈ match
Try
DeepSeek V4 Flash Vision
Fast vision-language model (experimental). Image understanding at Flash cost.
DeepSeek128K$0.308 / $1.231OpenRouter N/A
Exclusive
Try
MiniMax4 models
MiniMax-M3Flagship
MiniMax M3. Flagship with up to 512K context.
MiniMax512K$0.30 / $1.20$0.30 / $1.20
≈ match
Try
MiniMax-M2.7
MiniMax M2.7. Improved reasoning over M2.5.
MiniMax256K$0.30 / $1.20$0.25 / $1.00Try
MiniMax-M2.5
MiniMax M2.5. Solid all-round chat and content generation.
MiniMax256K$0.27 / $1.08$0.27 / $1.08
≈ match
Try
MiniMax-Text-01
256K ctx, strong Chinese text model.
MiniMax256K$0.14 / $1.10$0.20 / $1.10
30% cheaper input
Try
Moonshot6 models
Kimi-K3Flagship
Moonshot Kimi K3. Flagship 1M-context multimodal reasoning.
Moonshot1M$2.90 / $14.50$3.00 / $15.00
~3% cheaper
Try
Kimi-K2.8-Preview
Moonshot next-gen preview. Agentic reasoning & coding, 256K context.
Moonshot256K$0.93 / $3.89OpenRouter N/A
Exclusive
Try
Kimi-K2.7-Code
Kimi K2.7 Code. Coding-specialized, agentic dev workflows.
Moonshot256K$0.67 / $3.40$0.67 / $3.40
≈ match
Try
Kimi-K2.7-Code HighSpeed
Kimi K2.7 Code HighSpeed. Premium low-latency coding tier.
Moonshot256K$2.07 / $8.61OpenRouter N/A
Exclusive
Try
Kimi-K2.6
Kimi K2.6. Upgraded reasoning and coding.
Moonshot256K$0.95 / $4.00$0.95 / $4.00
≈ match
Try
Kimi-K2.5
Moonshot Kimi K2.5. Long-context reasoning specialist.
Moonshot256K$0.60 / $3.00$0.60 / $3.00
≈ match
Try
Tencent11 models
Hunyuan Hy4 PreviewFlagship
Tencent next-gen flagship preview. 1M context, strong CN & agentic reasoning.
Tencent1M$0.834 / $2.501$0.834 / $2.501
≈ match
Try
Hunyuan Hy3
Tencent Hunyuan flagship. Strong Chinese & general reasoning, agentic tool use.
Tencent256K$0.18 / $0.72$0.20 / $0.80
10% cheaper
Try
Hunyuan T1 Vision
HY-Vision-1.5-Thinking. Deep-reasoning vision-language: OCR, charts, docs.
Tencent$0.55 / $1.64OpenRouter N/A
Exclusive
Try
Hunyuan Vision 2.0 Instruct
Fast multimodal vision-language model. Image understanding, OCR, VQA.
Tencent$1.154 / $2.692OpenRouter N/A
Exclusive
Try
Hunyuan Turbos Vision Video
Vision-language model for image & video frame understanding.
Tencent$0.462 / $1.385OpenRouter N/A
Exclusive
Try
Hunyuan MT2 Pro
Dedicated machine-translation model. 30+ languages, domain-aware.
Tencent32K$0.29 / $0.87OpenRouter N/A
Exclusive
Try
Hunyuan MT2 Plus
Dedicated machine-translation model. 30+ languages, domain-aware.
Tencent8K$0.072 / $0.287OpenRouter N/A
Exclusive
Try
Hunyuan Role
Role-play & character model. Persona consistency, immersive dialogue.
Tencent32K$0.47 / $1.86OpenRouter N/A
Exclusive
Try
Hunyuan Role
Role-play & character model. Persona consistency, immersive dialogue.
Tencent32K$0.369 / $1.477OpenRouter N/A
Exclusive
Try
Hunyuan Hy3 Preview
Preview build of Hy3. Lowest-cost entry into the Hunyuan family.
Tencent256K$0.127 / $0.423$0.060 / $0.21Try
Hunyuan MT2 Lite
Budget machine-translation tier. Fast, low-cost, 30+ languages.
Tencent8K$0.043 / $0.172OpenRouter N/A
Exclusive
Try
Xiaomi1 models
MiMo-V2.5-ProFlagship
Xiaomi flagship reasoning model. 1M context, strong math & code.
Xiaomi1M$0.427 / $0.855$0.435 / $0.87
≈ match
Try
Zhipu AI11 models
GLM-5.3Flagship
New Zhipu flagship. Coding & agentic, 1M context.
Zhipu AI1M$1.25 / $4.00$1.40 / $4.40
11% cheaper
Try
GLM-5.2
Zhipu's newest flagship. Top reasoning & agentic coding.
Zhipu AI128K$1.19 / $3.74$1.19 / $3.74
≈ match
Try
GLM-5.1
Latest GLM iteration. Improved speed & accuracy.
Zhipu AI128K$1.05 / $3.96$1.26 / $3.96
input cheaper
Try
GLM-5
Zhipu flagship. Strong reasoning & tool calling.
Zhipu AI128K$0.60 / $1.92$0.60 / $1.92
≈ match
Try
GLM-5V-Turbo
Fast multimodal vision. Low-latency image understanding.
Zhipu AI128K$1.08 / $3.69$1.20 / $4.00
10% cheaper
Try
GLM-4.6V
Vision-language model. Image understanding & analysis.
Zhipu AI128K$0.30 / $0.90$0.30 / $0.90
≈ match
Try
GLM-5.3-Flash
Zhipu budget flagship. Coding & agentic at 1M context, lowest-cost tier.
Zhipu AI1M$0.08 / $0.28$0.075 / $0.25Try
GLM-5-Turbo
Fast GLM-5 variant. Low latency, high throughput for production.
Zhipu AI128K$1.08 / $3.69$1.20 / $4.00
8% cheaper
Try
GLM-4.7-Flash
FREE forever. 30B-class agentic coding & fast chat.
Zhipu AI202KFREE$0.06 / $0.40
free here
Try
GLM-4.6V-Flash
FREE forever. Vision model for basic image tasks.
Zhipu AI128KFREEOpenRouter N/ATry
GLM-4.1V-Thinking-Flash
FREE forever. Vision reasoning — step-by-step analysis of images.
Zhipu AI128KFREEOpenRouter N/ATry
Prices per 1M tokens (input / output). "% cheaper" = NovAI vs OpenRouter live list price (checked 2026). "N/A" = not offered on OpenRouter.
ModelVendorSpecsNovAIOpenRouter
Alibaba4 models
Qwen ImageFlagship
Flagship Qwen image generation. Photorealistic, highly detailed compositions.
Alibaba2K | T2I | Photoreal$0.0385 / imageOpenRouter N/A
Official direct
Try
Qwen Image Plus
Advanced text-to-image with precise prompt following and text rendering.
Alibaba2K | T2I | Text rendering$0.0308 / imageOpenRouter N/A
Official direct
Try
Wan 2.5 T2I Preview
Next-gen Wan image model (preview). Improved aesthetics and coherence.
Alibaba2K | T2I | Preview$0.0308 / imageOpenRouter N/A
Official direct
Try
Wan 2.2 T2I Flash
Fast text-to-image generation with high-quality outputs at low cost.
Alibaba2K | T2I | Fast$0.0215 / imageOpenRouter N/A
Official direct
Try
ByteDance4 modelsPro variants first; direct and TokenHub routes retain their original specifications.
Seedream 5.0 ProFlagship OFFICIAL DIRECT
Latest image gen. 4K, editing consistency, small-text rendering.
ByteDance4K | T2I | I2I | Multi-ref | Group photos$0.034 / image$0.04 / image
15% cheaper
Try
Seedream v5.0 Pro (TokenHub)
ByteDance Seedream via TokenHub gateway. Layer decomposition, transparent background.
ByteDance1K-2K | T2I | Ref ≤10 | Layersfrom $0.046 / imageOpenRouter N/A
Exclusive
Try
Seedream 5.0 OFFICIAL DIRECT EXCLUSIVE
Cost-effective 2K image gen. High-volume production.
ByteDance2K | T2I | I2I | Batch gen$0.027 / imageN/A
Exclusive
Try
Seedream v5.0 Lite (TokenHub)
Fast Seedream via TokenHub gateway. Group images, web-search tool, up to 4K.
ByteDance2K-4K | T2I | Ref ≤14 | Group$0.034 / imageOpenRouter N/A
Exclusive
Try
Tencent1 models
Hunyuan Image 3.0Flagship
Tencent flagship text-to-image via /v1/images/generations. Sync, prompt rewrite, custom watermark.
Tencent≤1024² | T2I | Ref 0-3 | Sync$0.031 / imageOpenRouter N/A
Exclusive
Try
Vidu (Shengshu)1 models
Vidu Image Q2
Vidu (Shengshu) image gen via TokenHub. Text-to-image, multi-reference, image editing.
Vidu (Shengshu)1080p-4K | T2I | Ref 0-7 | Editfrom $0.029 / imageOpenRouter N/A
Exclusive
Try
WAND3 modelsPublished model brand; developer attribution is not independently confirmed.
WAND Vega Image ProFlagship
WAND Vega top tier. Highest-fidelity image generation.
WAND1K-4K | T2I | Ref 0-6 | Layersfrom $0.146 / imageOpenRouter N/A
Exclusive
Try
WAND Vega Image Flash
WAND Vega mid tier. Balanced quality and speed for production.
WAND1K-4K | T2I | Ref 0-6from $0.069 / imageOpenRouter N/A
Exclusive
Try
WAND Vega Image Lite
WAND Vega entry tier. Fast, cost-effective text/reference image gen.
WAND1K-4K | T2I | Ref 0-3from $0.025 / imageOpenRouter N/A
Exclusive
Try
Zhipu AI1 models
CogView-3-Flash FREE FOREVER
Free image generation via /v1/images/generations. Zero cost, forever.
Zhipu AIT2I | 512-1024px | FreeFREEN/A
Exclusive
Try
TokenHub image models (Hunyuan / Vidu / Seedream / WAND Vega) are billed by actual upstream token usage; the per-image prices shown are reference values at base resolution and scale with resolution & reference-image count. Legacy flat-price models: Seedream 5.0, CogView-3-Flash.
ModelVendorSpecsNovAIOpenRouter
ByteDance3 models
Seedance 2.5Flagship OFFICIAL DIRECT
Flagship 2026 video. Native synchronized audio, cinematic quality.
ByteDanceUp to 30s | 480p/720p | T2V/I2V | Native audio480p $0.14 / 720p $0.32/s
Trial: 480p/5s, 1/day
Not yet listed
Exclusive launch
Try
Seedance 2.0 OFFICIAL DIRECT
Flagship video. Multi-modal, native audio, 4K cinematic.
ByteDance4-15s | 480p-4K | Text/Img/Vid/Audio | Native audio480p $0.067 / 720p $0.163 / 1080p $0.329 / 4K $1.00/s480p $0.067 / 720p $0.199 / 1080p $0.496 / 4K $1.36/s
720p+ 18-34% cheaper
Try
Seedance 2.0 Fast OFFICIAL DIRECT
Fast variant. ~30% faster, cost-optimized generation.
ByteDance4-15s | 480p-4K | Faster gen | Multi-modal480p $0.054 / 720p $0.131 / 1080p $0.264 / 4K $0.802/s480p $0.054 / 720p $0.121 / 1080p $0.302 / 4K $0.625/s
1080p 13% cheaper
Try
MiniMax4 models
Hailuo H3 MaxFlagship OFFICIAL DIRECT
MiniMax top-tier quality variant of H3 for text-to-video.
MiniMaxT2V | 768P/1080P | Async task$1.54 / 1M tok
≈ $0.077/s · $0.385/5s
N/A
Exclusive
Try
Hailuo H3 OFFICIAL DIRECT
MiniMax flagship text-to-video. High-fidelity cinematic motion.
MiniMaxT2V | 768P/1080P | Async task$1.54 / 1M tok
≈ $0.077/s · $0.385/5s
N/A
Exclusive
Try
Hailuo 2.3 OFFICIAL DIRECT
MiniMax text-to-video. Balanced quality and generation speed.
MiniMaxT2V | 768P/1080P | Async task$1.54 / 1M tok
≈ $0.062/s · $0.308/5s
N/A
Exclusive
Try
Hailuo 2.3 Fast OFFICIAL DIRECT
MiniMax image-to-video. Animate a still first frame into a clip.
MiniMaxI2V only | 768P/1080P | Async task$1.54 / 1M tok
≈ $0.042/s · $0.208/5s
N/A
Exclusive
Try
Tencent1 models
Hunyuan Video 1.5 OFFICIAL DIRECT
Tencent Hunyuan video gen. Text & image to video.
TencentT2V | I2V | 720p/1080p | Async task$1.54 / 1M tok
≈ $0.046/s · $0.231/5s
N/A
Exclusive
Try
Zhipu AI1 models
CogVideoX-Flash FREE FOREVER
Free video generation via /v1/video/generations. Zero cost, forever.
Zhipu AIT2V | 6s clips | 480p/720p | FreeFREEN/A
Exclusive
Try
Seedance & CogVideoX-Flash are billed per second of generated video (resolution-based). TokenHub video models (MiniMax Hailuo / Hunyuan Video) are billed by actual upstream token usage at $1.54/1M tokens; the per-second values are reference estimates — an estimate is pre-deducted at submit and reconciled to the real token count when the task succeeds.
ModelVendorModalityNovAI / 1M tokensOpenRouter / 1M tokens
Alibaba2 models
Qwen3.7 Text Embedding
High-quality multilingual text vectors for RAG, search and clustering.
AlibabaText · multilingual$0.11 inputOpenRouter N/A
Official direct
Try
Text Embedding V4
General-purpose text embeddings. Dense + sparse, multilingual retrieval.
AlibabaText · dense+sparse$0.08 inputOpenRouter N/A
Official direct
Try
Kinfra4 modelsPublished model brand; text and multimodal retrieval are different tasks. Larger variants lead within each family.
Kinfra Text Embedding 4B
High-quality 2560-dim text vectors for RAG, search & clustering.
KinfraText · 2560d$0.0923 inputOpenRouter N/A
Exclusive
Try
Kinfra VL Embedding 8B
Multimodal vision-language vectors across text, image & video.
KinfraMultimodal$0.0923 inputOpenRouter N/A
Exclusive
Try
Kinfra Text Embedding 0.6B
Fast, compact 1024-dim text vectors for high-volume retrieval.
KinfraText · 1024d$0.0769 inputOpenRouter N/A
Exclusive
Try
Kinfra VL Embedding 2B
Fast, compact multimodal vectors for cross-modal retrieval.
KinfraMultimodal$0.0769 inputOpenRouter N/A
Exclusive
Try
Billed per 1M input tokens via POST /v1/embeddings. Text models take a string or list of strings; multimodal (VL) models take an object array. Output (the vectors) is free.
ModelVendorTypeNovAI / 1M tokensOpenRouter
Alibaba3 models
Qwen TTS Latest
Latest Qwen TTS build with improved naturalness and voice stability.
AlibabaTTS · url$0.013 / 1K charsOpenRouter N/A
Official direct
Try
Qwen TTS
Natural multilingual text-to-speech with multiple voices and expressive prosody.
AlibabaTTS · url$0.013 / 1K charsOpenRouter N/A
Official direct
Try
Qwen3 TTS Flash
Fast, low-latency TTS for realtime voice and streaming applications.
AlibabaTTS · url$0.0145 / 1K charsOpenRouter N/A
Official direct
Try
MiniMax8 modelsSpeech synthesis: HD, then Turbo. Music generation is a separate task; no cross-task ranking.
MiniMax Speech 2.8 HD
Latest flagship TTS. Studio-grade voices, emotion control, 32kHz.
MiniMaxTTS · url/binary$1.54 / 1M tok
≈$0.054 / 1K chars
OpenRouter N/A
Exclusive
Try
MiniMax Speech 2.6 HD
High-definition 2.6 TTS. Natural prosody, multilingual voices.
MiniMaxTTS · url/binary$1.54 / 1M tok
≈$0.054 / 1K chars
OpenRouter N/A
Exclusive
Try
MiniMax Speech 02 HD
Speech-02 HD TTS. Rich, expressive timbres; clone-ready.
MiniMaxTTS · url/binary$1.54 / 1M tok
≈$0.054 / 1K chars
OpenRouter N/A
Exclusive
Try
MiniMax Speech 2.8 Turbo
Fast 2.8 TTS. Low-latency voice for realtime & streaming apps.
MiniMaxTTS · url/binary$1.54 / 1M tok
≈$0.031 / 1K chars
OpenRouter N/A
Exclusive
Try
MiniMax Speech 2.6 Turbo
Fast 2.6 TTS. Cost-optimized voice synthesis at scale.
MiniMaxTTS · url/binary$1.54 / 1M tok
≈$0.031 / 1K chars
OpenRouter N/A
Exclusive
Try
MiniMax Speech 02 Turbo
Speech-02 Turbo TTS. Quick, economical speech generation.
MiniMaxTTS · url/binary$1.54 / 1M tok
≈$0.031 / 1K chars
OpenRouter N/A
Exclusive
Try
MiniMax Music v3.0
Latest music gen. Full songs from lyrics + style prompt; vocal or instrumental.
MiniMaxMusic · url$1.54 / 1M tok
≈$0.154 / song
OpenRouter N/A
Exclusive
Try
MiniMax Music v2.6
Music generation. Lyrics-to-song with configurable style & structure.
MiniMaxMusic · url$1.54 / 1M tok
≈$0.154 / song
OpenRouter N/A
Exclusive
Try
Tencent1 models
Hunyuan ASR 3.0
Hunyuan speech-to-text. Accurate transcription, timestamps, SRT subtitles.
TencentASR · text$1.54 / 1M tok
≈$0.003 / min audio
OpenRouter N/A
Exclusive
Try
WAND1 modelsPublished model brand; developer attribution is not independently confirmed.
WAND ASR v1
WAND speech-to-text. Multilingual transcription with language detection + subtitles.
WANDASR · text$1.54 / 1M tok
≈$0.007 / min audio
OpenRouter N/A
Exclusive
Try
Audio is billed by actual upstream token usage at $1.54 / 1M tokens (10元/M upstream — the same rate as image & video). POST /v1/audio/speech returns an audio URL (response_format=url) or binary bytes (mp3/wav/opus/…); POST /v1/audio/transcriptions returns text. Per-unit references scale with text length (TTS), track length (music) or audio duration (ASR).
ModelVendorSpecsNovAIOpenRouter
Tencent5 modelsGeneral 3D generation, specialist geometry/motion, then Express; different tasks are not benchmarked against each other.
Hunyuan 3D 3.1 OFFICIAL DIRECT
Tencent Hunyuan 3D 3.1. Latest iteration, improved geometry & detail.
TencentT23D | I23D | obj/glb | Async task$0.369 / gen
flat, refund on fail
N/A
Exclusive
Try
Hunyuan 3D 3.0 OFFICIAL DIRECT
Tencent Hunyuan 3D 3.0. High-quality text/image-to-3D mesh generation.
TencentT23D | I23D | obj/glb | Async task$0.369 / gen
flat, refund on fail
N/A
Exclusive
Try
Hunyuan 3D PolyGen (Image) OFFICIAL DIRECT
Tencent Hunyuan PolyGen image-to-3D. Multi-format export (glb + obj + fbx) with preview.
TencentI23D | glb/obj/fbx | Async task$0.554 / gen
flat, refund on fail
N/A
Exclusive
Try
Hunyuan 3D Motion OFFICIAL DIRECT
Tencent Hunyuan 3D motion. Generates animated 3D motion output (fbx).
TencentMotion | fbx | Async task$0.185 / gen
flat, refund on fail
N/A
Exclusive
Try
Hunyuan 3D Express OFFICIAL DIRECT
Tencent Hunyuan fast text/image-to-3D. Quick drafts, low-cost mesh generation.
TencentT23D | I23D | obj/zip | Async task$0.277 / gen
flat, refund on fail
N/A
Exclusive
Try
Tripo (VAST)2 modelsSeparate product lines; no verified relative ranking between 3.1 and P1.
Tripo 3D 3.1 OFFICIAL DIRECT
Tripo (VAST) 3D 3.1. Text/image-to-3D generation, glb output with rendered preview.
Tripo (VAST)T23D | I23D | glb | Async task$0.323 / gen
flat, refund on fail
N/A
Exclusive
Try
Tripo 3D P1 OFFICIAL DIRECT
Tripo (VAST) 3D P1. Text/image-to-3D generation, glb output with rendered preview.
Tripo (VAST)T23D | I23D | glb | Async task$0.323 / gen
flat, refund on fail
N/A
Exclusive
Try
3D models are billed at a flat per-generation price (default tier). The full price is pre-deducted at submit and refunded in full if the job fails or is cancelled — you only pay for successful generations. POST /v1/three_d/generations returns a task id (status queued); poll GET /v1/three_d/generations/{id} until succeeded to get downloadable model URL(s) (glb / obj / fbx / zip). Only the default generation mode is exposed, so the flat price always matches the cost.