Every model. One transparent price list.
Mina Labs runs 50+ AI models for images, video, music and 3D in one workspace, including full AI songs with vocals. You pay in credits, per generation, no subscription required. Failed generations are refunded automatically.
🖼️ Image models
Credits per generated image. All available in the Image, Logo, Banner, Human, Animal and Product studios.
| Model | Credits | Best for |
|---|---|---|
| Bria RMBG 2.0 Background Remover | FREE | Transparent-background PNG with clean hair and fine edges, licensed data, safe for commercial use |
| BiRefNet Background Remover | FREE | Clean transparent-background PNG from any photo |
| Bria Product Shot | FREE | Puts your product in any scene, white studio, marble, cafe table, without altering the product itself |
| Bria Product Dimensions | 2 | Draws measurement arrows (height, width, units) on your product photo |
| Qwen Image 3 Pro | 2 | Strong prompt following and clean text in images, the cheapest pro image model here |
| Grok Imagine Image | 3 | Fast, budget-friendly generations and edits |
| Seedream V5 Pro | 3 | High-quality image editing |
| Nano Banana / Nano Banana 2 | 4 | Photorealistic scenes, faces, product shots |
| Flux 2 Flash / Flux.1 Dev | 4 | Fast concept art and drafts |
| Grok Imagine Image Pro | 4 | Higher-quality Grok generations |
| GPT Image 2 | 1–5 | Prompt-faithful images, text in images (quality tiers) |
| Recraft V4.1 Pro Vector | 5 | Vector SVG logos, icons and illustrations |
| Ideogram Auto | 5 | Design-focused graphics with typography |
| Trellis 2 (Image to 3D) | 5–7 | Turn any image into an interactive 3D GLB model |
| Hunyuan 3D Pro (Image to 3D) | 15 | High-detail textured 3D GLB from a single image |
| Hunyuan 3D Pro (Text to 3D) | 15 | High-detail textured 3D GLB from a text prompt, no image needed |
| Nano Banana Pro | 6 | Premium photorealism |
| Flux 2 | 6 | Balanced quality / cost all-rounder |
| Gemini 2.5 Flash Image | 6 | Fast Google-quality images |
| Gemini 3 Pro Image | 8 | Top-tier Google image quality |
| Flux 2 Pro | 8 | Professional detail and lighting |
| Flux 2 Max | 10 | Maximum Flux quality |
| Krea 2 Turbo Style | 2–8 | Style transfer from your reference images (by resolution) |
🎬 Video models
Credits per clip at 720p (5 / 8 / 10 / 15 seconds where supported). 480p economy pricing is lower on supported models.
| Model | 5s | 8s | 10s | 15s |
|---|---|---|---|---|
| LTX 2.3 Distilled (fastest, budget) | 6 | 9 | 11 | 16 |
| LTX 2.3 22B | 7 | 12 | 14 | 21 |
| LTX 2.3 Quality | 11 | 17 | 21 | 32 |
| LTX 2.3 Distilled Styled (LoRA styles) | 7 | 10 | 13 | 19 |
| LTX 2.3 22B Styled (LoRA styles) | 8 | 13 | 16 | 24 |
| LTX 2.3 Fast (1080p / 1440p / 4K) | 1080p: 10–32 • 1440p: 20–64 • 4K: 38–126 credits by clip length (6–20s) | |||
| Kling 2.5 Turbo Pro | 14 | 18 | 22 | 31 |
| Gemini Omni Flash | 17 | 28 | 33 | ... |
| Gemini Omni Flash Image-to-Video | 27 | 42 | 53 | ... |
| Gemini Omni Flash Edit (AI video editing) | 26 | 41 | 51 | ... |
| Grok Imagine Video | 24 | 41 | 48 | 71 |
| Grok Imagine Image-to-Video (animate a photo) | 8 | 12 | 15 | 22 |
| Kling v3 Standard | 13 | 21 | 26 | 39 |
| Kling v3 Pro | 22 | 35 | 44 | 62 |
| Kling v3 Turbo Standard | 18 | 28 | 35 | 53 |
| Kling v3 Turbo Pro | 22 | 35 | 44 | 66 |
| Seedance 2.0 Mini | 22 | 35 | 44 | 66 |
| Seedance 2.0 Fast | 35 | 58 | 70 | 106 |
| Seedance 2.0 Premium | 44 | 70 | 88 | 132 |
| Seedance 2.5 (text / image / reference to video, clips up to 30s, native audio) | 720p: 5s 74 • 10s 148 • 15s 222 • 30s 444, 480p about half (35 / 69 / 104 / 207). One model, three modes, single-shot clips from 4 to 30 seconds | |||
| Flux 3 (Black Forest Labs, text, image, first-last frame, keyframes, extend) | 720p: 5s 27 • 10s 54 • 15s 80 • 20s 107, 1080p: 46 / 91 / 136 / 182. Extend Video costs more: ≈13 credits/sec at 720p, ≈17/sec at 1080p | |||
| Cosmos 3 Super (NVIDIA, animates a photo) | 2s: 4 • 3s: 6 • 4s: 8 • 5s: 10 • 6s: 12 • 7s: 14 credits, same price at every resolution, 7 seconds max | |||
| Veo 3.1 (4/6/8s, audio optional) | 32–188 by length, resolution (up to 4K) and audio, the Fast tier halves it (from 16) | |||
| Sync Lipsync v3 (talking avatar) | 3.5 credits per second of audio (35 per 10s) | |||
| Grok Imagine Video 1.5 Text to Video | A prompt on its own, no upload; 12 credits for 5s at 480p, 20 at 720p, 36 at 1080p, 1 to 15 seconds, billed per second | |||
| Grok Imagine Video 1.5 Image to Video | 12 credits for 5s at 480p, 20 at 720p, 36 at 1080p, 1 to 15 seconds, billed per second | |||
| Grok Imagine Video 1.5 Reference to Video | Up to 7 reference images cited in the prompt as <IMAGE_0>; 12 credits for 5s at 480p, 20 at 720p, 1 to 15 seconds | |||
| Mirage Avatar X: Script to Video | 9 credits per second of speech : the length comes from your script, about 4 seconds per 50 characters | |||
| Mirage Avatar X: Reference to Video | 9 credits per second of your audio (3–180s), driving a face from your video, your image, or a stock avatar | |||
| LTX 2.3 Reframe (change aspect ratio) | 4 credits/sec at 720p, 8 credits/sec at 1080p, billed on your input video length | |||
| MiniMax H3 (text / image / reference to video, 2K) | 5s: 41 • 8s: 65 • 10s: 82 • 15s: 122 credits, 2K throughout. Reference to Video includes 5 reference images; extra images and reference video add to the price shown before you generate | |||
| Kling 4K (image to video, native 4K) | 3s: 35 • 5s: 58 • 8s: 92 • 10s: 115 • 15s: 173 credits, billed per second, 4K straight out of the model with no upscale pass | |||
| Kling 4K Edit (rewrite a clip you already have) | Same per-second price as Kling 4K above, billed on the length of the video you upload; reference images are cited in the prompt as @Image1 | |||
| MiniMax H3 Max Turbo (text / image to video, 480p or 768p) | Launch price while the model's promotion lasts, to 7 September: 768p 5s 2 • 10s 3 • 15s 5, 480p 5s 1 • 10s 2 • 15s 3. From 7 September: 768p 5s 6 • 10s 11 • 15s 17, 480p 5s 4 • 10s 7 • 15s 11. Billed per second, any whole 5 to 15 seconds | |||
| Pixverse C1 (text / image / reference / transition) | Per second by resolution + audio, a 5s clip runs 6 (360p) to 24 (1080p + audio); scales with length | |||
Kling v3 Standard, v3 Pro and 2.5 Turbo Pro also animate an uploaded photo, image-to-video costs the same per second as the text tier above.
Seedance 2.0 also animates an uploaded photo, Mini, Fast and Premium image-to-video cost the same per second as the text-to-video tier above, and take an optional last frame.
Kling v3 Turbo is one model in the studio, pick Standard or Pro, and Text to Video (prompt only) or Image to Video (animate an uploaded photo). The price updates to the option you choose; image-to-video costs the same per tier as text-to-video.
🎵 Music
| Model | 30s | 60s | 120s | 180s |
|---|---|---|---|---|
| CassetteAI Music Generator | 3 | 5 | 10 | 15 |
| GPT Audio, AI Voice | 4 | Spoken audio from text, scripts, answers, narration (OpenAI voices) | ||
| ElevenLabs Music | 32 | 32 | 64 | 96 |
| MiniMax Music 2.6full songs with vocals | 6 | 6 | 6 | 6 |
ElevenLabs Music is billed per output minute (rounded up), so 30s and 60s tracks both cost one minute. MiniMax Music 2.6 generates a complete track (singing, backing music and arrangement) at a flat 6 credits.
🧪 Coming soon
Already integrated and being priced, generation unlocks shortly.
| Model | What it will do |
|---|---|
| Reve 2.1 / Remix / Edit | Text-to-image, image remixing and precise editing |
| VEED Lipsync V2 | Alternative talking-avatar lipsync engine |
| Sonilo V1.1 | Video-to-video transformation |
❓ Frequently asked questions
How much does it cost to generate an AI image?
AI images start at 2 credits (Qwen Image 3 Pro) up to 10 for Flux 2 Max. Background removal is free, and pay per generation, no subscription, you only pay per generation, no subscription.
Can I generate AI music with vocals?
Yes. MiniMax Music 2.6 creates complete songs with singing, backing music and arrangements for a flat 6 credits. ElevenLabs Music and CassetteAI cover instrumental and vocal tracks priced by length.
What is the cheapest AI video generator?
LTX 2.3 Distilled at 6 credits for a 5-second clip. Kling v3 Standard starts at 13, and Veo 3.1 with native audio scales up to 4K. Every price shows on the Generate button before you spend.
Is there a free AI model?
Yes, Bria RMBG 2.0 and the BiRefNet Background Remover both return a clean transparent-background PNG at no credit cost. The paid models are pay-per-generation, with no subscription.
Do I need a subscription?
No. Mina Labs charges per generation in credits with no monthly fee. Failed generations refund automatically, and 1 credit costs roughly $0.064–$0.09 depending on the pack.
💳 Credit packs
| Pack | Credits | Price |
|---|---|---|
| Starter | 100 | $9 |
| Creator | 200 | $15 |
| Business | 250 | $18 |
| Pro | 500 | $31, best value |
⚖️ Model comparisons
Not sure which model to pick? These guides compare them head-to-head.