Every model, one account
Makemancer gives Claude, ChatGPT and Codex access to 80 media models through one connector: image models such as Nano Banana and GPT Image, video models such as Seedance and Wan, ElevenLabs voices, music and upscaling. The assistant picks a default for each job; name a model and it uses that one.
Model guides
Each guide shows our own results from the model, its prices and how to ask for it in Claude or ChatGPT.
- Nano BananaUse Google's Nano Banana 2, Nano Banana Pro and Nano Banana 2 Lite from Claude, ChatGPT or Codex. Real results, prompts to copy and the price of each image.
- GPT ImageUse OpenAI's GPT Image 2, GPT Image 2.5 Flare and GPT Image 2.5 Sunburst from Claude, ChatGPT or Codex. Our own results, prompts to copy and the price of each image.
- SeedreamUse ByteDance's Seedream 5 Pro and Seedream 5 Lite from Claude, ChatGPT or Codex. Our own results, prompts to copy and the price of each image.
- FLUXUse Black Forest Labs' FLUX 2 Pro, FLUX Kontext Max and FLUX 3 video from Claude, ChatGPT or Codex. Our own results, prompts to copy and the price of each result.
- Grok ImagineUse xAI's Grok Imagine Image 2.0 and Grok Imagine 1.5 video from Claude, ChatGPT or Codex. Our own results, prompts to copy and the price of each result.
- SeedanceUse ByteDance's Seedance 2.0 Mini and Seedance 2.5 from Claude, ChatGPT or Codex. Our own clips, prompts to copy and the price of each video.
- MiniMax and HailuoUse MiniMax H3, H3 Max, H3 Max Turbo, H3 Max Recast, Hailuo-02 Pro and Hailuo 2.3 Fast from Claude, ChatGPT or Codex. Our own clip, prompts to copy and the price of each video.
- WanUse Alibaba's Wan 3.0 from Claude, ChatGPT or Codex: video from text, frames, references or a document or webpage brief. Our own clips, prompts to copy and the price of each video.
Images
Nano Banana
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Nano Banana 2image/nano-banana-2 | Google Nano Banana 2. Default for image generation and targeted edits, with strong instruction following and reference consistency at 1K, 2K and 4K. | 1K · image | 12$0.12 |
| Nano Banana Proimage/nano-banana-pro | Google Gemini 3 Pro image model, for when it is selected by name. Up to 8 reference images, 1K to 4K. | 4K · image | 30$0.30 |
| Nano Banana 2 Liteimage/nano-banana-lite | Google Gemini 3.1 Flash Lite Image (Nano Banana Lite). Image generation and editing at 1K resolution. | per image | 5$0.05 |
GPT Image
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| GPT Image 2image/gpt-image-2 | OpenAI GPT Image 2. Fast, high-quality generation and editing with strong typography and prompt adherence. | 1K · image | 47$0.47 |
| GPT Image 2.5 Flareimage/gpt-image-2.5-flare | OpenAI GPT Image 2.5 Flare. Fast, high-quality everyday image generation and editing. | 1K · image | 12$0.12 |
| GPT Image 2.5 Sunburstimage/gpt-image-2.5-sunburst | OpenAI GPT Image 2.5 Sunburst. Precision-focused generation and editing with strong instruction following, typography, and detail fidelity. | 1K · image | 12$0.12 |
| GPT Image 2.6 Flareimage/gpt-image-2.6-flare | OpenAI GPT Image 2.6 Flare. OpenAI's fastest high-quality image generation and editing. | 1K · image | 9$0.09 |
| GPT Image 2.6 Sunburstimage/gpt-image-2.6-sunburst | OpenAI GPT Image 2.6 Sunburst. Precision-focused generation and editing that keeps details consistent from one edit to the next, with strong instruction following and typography. | 1K · image | 9$0.09 |
Seedream
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Seedream 5 Proimage/seedream-5-pro | ByteDance Seedream 5.0 Pro. Product and marketing generation and reference-guided editing at 1K or 2K. | 2K · image | 18$0.18 |
| Seedream 5 Liteimage/seedream-5-lite | ByteDance Seedream 5 Lite. Reasoning-first 2K/3K/4K generation and precise edits with up to 14 reference images. | 2K · image | 7$0.07 |
FLUX
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| FLUX 2 Proimage/flux-2-pro | FLUX.2 Pro. Zero-config professional quality at 1K or 2K, no parameter tuning needed. | 1K · image | 7$0.07 |
| FLUX Kontext Maximage/flux-kontext-max | FLUX Kontext Max for consistent, precise edits and typography. Edit only: requires image_url. | per image | 16$0.16 |
Grok Imagine
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Grok Imagine Image 2.0image/grok-imagine-2.0 | xAI Grok Imagine Image 2.0. Text generation and natural-language editing with up to 5 source images. | per image | 12$0.12 |
Recraft
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Recraft V3image/recraft-v3 | Recraft V3 for design and illustration in raster styles: realistic_image, digital_illustration and their substyles. | per image | 8$0.08 |
More image models
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Ideogram 4.5image/ideogram-4.5 | Ideogram 4.5. Typography and image generation, consistent multi-turn editing with up to 5 ordered source/reference images (image_url first), and optional masked edits (mask_image_url: black is edited, white kept). | per image | 6$0.06 |
| Bria Fibo 1.5image/bria-fibo-1.5 | Bria's licensed-data image family for controllable generation, strong typography and product imagery, plus instruction-based edits with up to 4 ordered references (Image 1 is image_url). | 1MP · image | 8$0.08 |
Video
FLUX
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| FLUX 3video/flux-3 | Black Forest Labs FLUX 3. Native-audio 5-20 second video at 720p/1080p from text or an opening keyframe.Opens after your first credit purchase | 720p · 5 s | 170$1.70 |
| FLUX 3 Maxvideo/flux-3-max | A post-trained version of Black Forest Labs' FLUX 3 video model. Native-audio 5 to 20 second video at 720p/1080p from text, first/last frames, or up to 10 timed keyframes (extras.keyframes). | 720p · 5 s | 101$1.01 |
Grok Imagine
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Grok Imagine 1.5video/grok-imagine-1.5 | xAI Grok Imagine 1.5 with native audio, from one image (first frame or reference), 1-15 s at 480p/720p.Opens after your first credit purchase | 480p · 8 s | 137$1.37 |
Seedance
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Seedance 2.0 Minivideo/seedance-2.0-mini | ByteDance Seedance 2.0 Mini, the fastest and lowest-cost Seedance tier. Native audio, image/video/audio references, 480p/720p, 4-15 s. | 480p · 5 s | 15$0.15 |
| Seedance 2.5video/seedance-2.5 | Premium cinematic video with native audio, 4-30 second clips, first/last frames, and up to 30 image, 10 video and 10 audio references. draft=true makes a 480p draft; for the final cut, run the request again without draft.Opens after your first credit purchase | 480p · 5 s | 93$0.93 |
Kling
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Kling 4.0video/kling-4.0-omni | Kling 4.0 with native audio, 3-30 s at 720p/1080p: first/middle/last frames, or up to 10 reference images, 7 image elements (extras.contents) and 5 reference videos (video_urls, 3-30 s each, 30 s combined). Cite references as @Image1, @Video1 and elements as @id. extras.video_mode=edit makes video_urls[0] the source to edit. Video inputs are charged on output plus the longer of output and input. Elements in contents are built from your own photos; describe voices in the prompt.Opens after your first credit purchase | 720p · 5 s | 157$1.57 |
| Kling 3.0 Motion Controlvideo/kling-3.0-motion-control-pro | Transfers a recorded performance from one driving clip (3-30 s, 3-10 s with image orientation) to the person in a prepared scene image. Output length follows the driving clip. extras.elements (one face binding) and extras.keep_original_sound run on a second provider.Opens after your first credit purchase | 10 s video input | 336$3.36 |
MiniMax and Hailuo
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| MiniMax H3video/minimax-h3 | MiniMax H3 (not H3 Max). 768p or 2K output, native audio, first/last frames, and image/video/audio references, 4-15 s. | 768p · 4 s | 64$0.64 |
| MiniMax H3 Maxvideo/minimax-h3-max | Post-trained MiniMax H3 for very fast 480p/768p/1080p generation with strong prompt adherence, native audio, optional first/last frames, and multimodal references. | 480p · 5 s | 50$0.50 |
| MiniMax H3 Max Turbovideo/minimax-h3-max-turbo | Cheap, fast drafts. Throughput-optimized MiniMax H3 Max for text and literal first/last-frame generation at 480p/768p/1080p plus multimodal reference-to-video. | 480p · 5 s | 25$0.25 |
| H3 Max Recastvideo/minimax-h3-max-recast | Recast people in existing footage (5-30 s, no shot over 15 s) from 1 to 4 ordered photos; aims to keep motion, camera, cuts and audio. 768P or 1080P. | 8 s video input | 720$7.20 |
| H3 Max Extend Videovideo/minimax-h3-max-extend | Extend one existing video (video_urls, 1.625 to 60 s, up to 50 MB) with 5 to 15 seconds of new footage that keeps its characters, setting, motion and style. Charged per generated second plus the source video as reference tokens. | 5 s video input | 193$1.93 |
| H3 Max Video Insertvideo/minimax-h3-max-insert | Replace the span between extras.start_time and extras.resume_time of one source clip (video_urls) with a generated 5 to 13 second scene, then resume the original footage. Optional image_urls and extras.reference_video_urls guide it (Image 1, Video 1). 480p or 768p. | 768p · 5 s | 90$0.90 |
| H3 Max Lip Syncvideo/minimax-h3-max-lip-sync | Animate one still face image (image_url, the literal first frame) with a finished speech recording (audio_url, 5 seconds to 15 minutes); the video runs as long as the audio and keeps it. 480p to 2K, without a prompt. Works best with human faces. | 8 s audio input | 80$0.80 |
| Hailuo-02 Provideo/hailuo-02-pro | Great physics, director-level camera controls. Text or first/last-frame animation, 6 s at 1080p.Opens after your first credit purchase | per request | 80$0.80 |
| Hailuo 2.3 Fastvideo/hailuo-2.3-fast | Hailuo 2.3 Fast tier for animating a reference image, 6 or 10 s.Opens after your first credit purchase | 6 s | 23$0.23 |
Wan
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Wan 3.0video/wan-3.0 | Alibaba Wan 3.0. Native-audio video from text, first/last frames, up to 10 image, 5 video and 5 audio references, or one document (file_url) or webpage (web_url) brief, 2-30 s at 480p to 1080p.Opens after your first credit purchase | 480p · 5 s | 33$0.33 |
Gemini Omni
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Gemini Omni 1.1 Flashvideo/gemini-omni-flash | Google Gemini Omni 1.1 Flash. Default video model: fast multimodal generation and editing from text, images, and one source video, with native audio, at 720p, 1080p or 4k (4k at 10 seconds). | 720p · 4 s | 40$0.40 |
More video models
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| HappyHorse 1.1video/happyhorse-1.1 | Alibaba HappyHorse 1.1. Joint audio and video in one pass with native lip-sync, strong character consistency from 1-9 reference images (image_urls, cited as [Image 1]), or a first frame. 3-15 s at 720p/1080p.Opens after your first credit purchase | 720p · 5 s | 98$0.98 |
| HappyHorse 1.0video/happyhorse-1.0 | Alibaba HappyHorse 1.0. Joint audio and video in one pass with native lip-sync, from text or a first frame, 3-15 s at 720p/1080p.Opens after your first credit purchase | 720p · 5 s | 98$0.98 |
| LTX 2.3 Fastvideo/ltx-2.3-fast | Lightricks LTX 2.3 Fast. Native-audio text or image video, 6-20 s at 1080p.Opens after your first credit purchase | 1080p · 6 s | 72$0.72 |
| HeyGen Video 1.0video/heygen-video-1 | HeyGen's post-trained MiniMax H3 video model. Text, a literal first frame (image_url), or ordered image/video/audio references (cite <Picture 1>, <Video 1>, <Audio 1>) to a 5 to 15 second clip with native audio at 480p or 768p. Reference mode costs more and bills reference video seconds too.Opens after your first credit purchase | 768p · 5 s | 30$0.30 |
| HeyGen Avatar 4video/heygen-avatar-4 | Animate a face photo (image_url) with a script in prompt and a HeyGen voice, or with exact speech in audio_urls. 720p or 1080p.Opens after your first credit purchase | per input | 120$1.20 |
| Sync 3 Lip Syncvideo/sync-3 | Lip-sync one finished video (video_urls) to one prepared speech file (audio_urls). Charged per second of the source video; longer audio is cut off.Opens after your first credit purchase | 12 s video input | 324$3.24 |
| PixVerse Lip Syncvideo/pixverse-lipsync | Lip-sync one finished video (video_urls) to one finished speech recording (audio_urls); the mouth closes during pauses. Charged per second of the source video. Animals and cartoons can fail or distort.Opens after your first credit purchase | 12 s video input | 120$1.20 |
| DreamActor v2video/dreamactor-v2 | ByteDance DreamActor v2. Transfer full-body motion, facial expressions, and lip movement from a driving video (up to 30 s) to a person, illustration, or pet.Opens after your first credit purchase | 10 s video input | 100$1 |
Voice, music and sound
ElevenLabs
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| ElevenLabs v4speech/elevenlabs-v4 | Default speech model: stock voices, narration, expressive delivery with audio tags, 90+ languages. | per 1,000 characters | 20$0.20 |
| ElevenLabs Multilingual v2speech/elevenlabs-v2 | Stable professional narration, 29 languages, no square-bracket audio tags. Voices are ElevenLabs voice IDs. | per 1,000 characters | 20$0.20 |
| ElevenLabs Sound Effects v2sfx/elevenlabs-sound-v2 | Prompted sound effects, Foley, impacts, ambience, and seamless loops up to 30 seconds. Without duration_seconds the model picks the length and the charge follows the delivered audio. | 3 s | 2$0.02 |
More voices
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Gemini 3.8 Flash TTSspeech/gemini-3.8-flash-tts | Google's expressive multilingual text-to-speech with exact transcript, voice selection, style control and two-speaker dialogue. | per 1,000 characters | 18$0.18 |
| Gemini 3.8 Flash-Lite TTSspeech/gemini-3.8-flash-lite-tts | Google's lighter multilingual text-to-speech with exact transcript, voice selection and style control. | per 1,000 characters | 12$0.12 |
| MiniMax Speech 2.8 HDspeech/minimax-hd | MiniMax Speech 2.8 HD. 300+ voices, emotion/speed/pitch/volume control, 40 languages. | per 1,000 characters | 17$0.17 |
| Chatterbox Englishspeech/chatterbox | English-only expressive speech and one-off voice matching from reference audio (audio_urls). | per 1,000 characters | 5$0.05 |
| Chatterbox Multilingualspeech/chatterbox-multilingual | Short multilingual speech and one-off cross-language voice matching in 23 languages, at most 300 characters. voice_id is the target language; with audio_url, give custom_audio_language instead. | per request | 3$0.03 |
| Orpheus TTSspeech/orpheus | Expressive speech with emotive tags like <laugh>, <sigh>, <gasp>. | per 1,000 characters | 10$0.10 |
| Kokorospeech/kokoro | Kokoro. Fast American English speech with clean output. | per 1,000 characters | 4$0.04 |
| Fish Audio Drama 3speech/fish-drama-3 | Expressive speech with [natural-language performance directions], paired delivery tags, multi-speaker extras.dialogue, and private cloned or designed Character voices. | per 1,000 UTF-8 bytes | 10$0.10 |
More sound effects
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Mirelo SFX 2.0sfx/mirelo-sfx-2.0 | Text sound effects (prompt) or a soundtrack for a video (video_url, audio only, existing speech kept by default), up to 600 seconds. Charged per output second per sample. | 8 s | 16$0.16 |
Music
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Musicmusic/standard | Songs and instrumentals. Give a prompt (description mode), or style + title with lyrics (custom mode). Returns two variations. | per generation (two tracks) | 16$0.16 |
| Music Extendmusic/standard-extend | Continue a generated track (audio_id from a music result tracks entry, optional task_id = its job_id) from continue_at seconds. Returns two variations. | per generation (two tracks) | 16$0.16 |
| Music Covermusic/standard-cover | Cover a song in a new style, keeping or replacing its lyrics: an uploaded song (audio_url, up to 8 minutes) or a generated track (audio_id from a music result, optional task_id = its job_id). Returns two variations. | per generation (two tracks) | 16$0.16 |
| Music Region Editmusic/standard-region-edit | Replace the lyrics of one time range (infill_start_s to infill_end_s) of a generated track (audio_id from a music result, optional task_id = its job_id). full_lyrics is the whole song with the new section in place. | per generation (two tracks) | 20$0.20 |
Upscaling, frame rate and cleanup
Upscaling
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| AuraSRupscale/aura-sr | Default image upscaler: fast, cheap 4x upscale. Source at most 3072 pixels on each side. | per input | 8$0.08 |
| Recraft Crisp Upscaleupscale/recraft-crisp | Crisp, clean image upscale; the provider picks the output size. Source at most 4096 pixels on each side. | per request | 2$0.02 |
| Topaz Image Upscaleupscale/topaz-image | Topaz image upscale by 1, 2 or 4x. | per input | 16$0.16 |
| Clarity Upscalerupscale/clarity | Most detailed image upscale, 1-4x including fractions, optionally prompt-guided. Output up to about 22 megapixels. | per input | 80$0.80 |
| ByteDance Video Upscalerupscale/bytedance | Default video upscaler: AI enhancement to 1080p, 2K, 4K, 6K or 8K at 24-120 fps, with content presets including one for AI-generated video (aigc). Pro tier for cinematic restoration and 10/12-bit output. Charged per second of source video by output size, frame rate and tier.Opens after your first credit purchase | 5 s video input | 8$0.08 |
| Topaz Video Upscaleupscale/topaz-video | Topaz video upscale by 1, 2 or 4x, source up to 2 minutes. Charged per second of source video by output size and frame rate.Opens after your first credit purchase | 10 s video input | 160$1.60 |
| RealESRGAN Videoupscale/realesrgan | Plain 1-8x video enlargement without generative detail. At most 1,000 source frames. Charged per output megapixel of every frame.Opens after your first credit purchase | 5 s video input | 89$0.89 |
| Topaz Starlight Precise 2.6upscale/topaz-starlight-precise-2.6 | Topaz Starlight Precise 2.6 generative video upscaler, 1-4x with output up to 1080p tall. Charged per second of source video.Opens after your first credit purchase | 5 s video input | 120$1.20 |
| Topaz Starlight HQupscale/topaz-starlight-hq | Topaz Starlight HQ generative video upscaler, 1-4x with output up to 1080p tall. Charged per second of source video, at least 24 credits.Opens after your first credit purchase | 5 s video input | 120$1.20 |
| Topaz Starlight Miniupscale/topaz-starlight-mini | Topaz Starlight Mini generative video upscaler, 1-4x with output up to 1080p tall. Charged per second of source video, at least 24 credits.Opens after your first credit purchase | 5 s video input | 120$1.20 |
| Topaz Starlight Fast 2upscale/topaz-starlight-fast-2 | Topaz Starlight Fast 2 generative video upscaler, 1-4x with output up to 1080p tall. Charged per second of source video, at least 24 credits.Opens after your first credit purchase | 5 s video input | 60$0.60 |
Background removal
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| BiRefNet v2background-removal/birefnet | Default image background remover: a transparent PNG. | per request | 2$0.02 |
| Recraft Remove Backgroundbackground-removal/recraft | Image background removal to a transparent PNG. Source 256-4096 px per side, at most 16 MP and 5 MB. | per request | 2$0.02 |
| Bria V-RMBG 3.0background-removal/bria-v3 | Video background removal with steady edges on solid subjects. Outputs webm, mp4 or ProRes. Charged per second of source video. | 10 s video input | 100$1 |
| BiRefNet Matting (video)background-removal/birefnet-matting | Video background removal with soft edges on see-through subjects; can miss the subject in busy or low-contrast shots. Outputs webm or mp4. Charged per second of source video. | 5 s video input | 21$0.21 |
| BEN v2 (video)background-removal/ben-v2 | Per-frame video background removal with hard edges. Outputs webm or mp4. Charged per second of source video. | 5 s video input | 50$0.50 |
| VEED Background Removalbackground-removal/veed | Video background removal that keeps semi-transparent subjects partly transparent. Outputs webm. Sources up to 30 fps. | 10 s video input | 36$0.36 |
| Bria V-RMBG 3.0 Green Screenbackground-removal/bria-v3-green-screen | For footage shot on green: removes the green and pulls green spill off the subject's edges. Outputs webm, mp4 or ProRes. Charged per second of source video. | 10 s video input | 100$1 |
| VEED Green Screenbackground-removal/veed-green-screen | For footage shot on green, with adjustable spill suppression. Outputs webm. Sources up to 30 fps. | 10 s video input | 40$0.40 |
Frame interpolation
| Model | Best for | Priced for | Credits |
|---|---|---|---|
| Topaz Apolloframe-interpolation/topaz-apollo | Topaz's general-purpose frame interpolation; a short clip takes about a minute. Up to 120 fps, or slow motion with slowdown_factor 2-8, for sources up to 1080p and 120 seconds. Charged per generated frame, at least 24 credits; in slow motion every output frame counts. | 5 s video input | 30$0.30 |
| Topaz Chronosframe-interpolation/topaz-chronos | Topaz frame interpolation tuned for real camera footage; about as fast as Apollo. Up to 120 fps, or slow motion with slowdown_factor 2-8, for sources up to 1080p and 120 seconds. Charged per generated frame, at least 24 credits; in slow motion every output frame counts. | 5 s video input | 30$0.30 |
| Topaz Aionframe-interpolation/topaz-aion | Topaz frame interpolation for extreme, complex motion such as particles, cloth and whip pans. Much slower: about half an hour for 5 seconds near 1080p. Up to 120 fps, or slow motion with slowdown_factor 2-8, for sources up to 1080p and 60 seconds. Charged per generated frame, at least 24 credits; in slow motion every output frame counts. | 5 s video input | 50$0.50 |
Questions
Which model does the assistant use if I do not name one?
Nano Banana 2 for images, Gemini Omni Flash for video and ElevenLabs v4 for speech, unless the request calls for something else, such as legible text in an image or a quick low-cost video draft. The assistant says which model it chose and what it costs before it runs.
Can I compare two models on the same prompt?
Yes. Ask for the same request on two or three models; each run is charged separately. The comparison pages show what that looks like.
Why do prices differ so much between models?
Each model has its own cost per image or per second of video, and resolution and length multiply it. The price list shows a sample for every model, and the assistant can quote any request first.
Are new models added?
Yes. When a model is added, it shows up here and in the assistant's model list without anything to update on your side.