The AI models behind Soloreach
18 models from 13 labs, each used for the step where it does best. Here's what each one does, what it costs in credits and when to pick it.
- Anthropic
- OpenAI
- Mistral AI
- DeepSeek
- Meta
- Alibaba
- ElevenLabs
- HeyGen
- Google DeepMind
- Kuaishou
- ByteDance
- MiniMax
Claude Sonnet 5.5
Anthropic · Text
The default model for hooks, scripts and articles.
QualitySpeed4 cr per scriptGrowth and upClaude Opus 5.5
Anthropic · Text
For long analyses, pillar articles and content plans.
QualitySpeed8 cr per scriptGrowth and upGPT-5.5
OpenAI · Text
Versatile, strong on structure and storytelling.
QualitySpeed7 cr per scriptGrowth and upGemini 3.1 Pro
Google · Text
Watches competitor videos and reads their pages.
QualitySpeedIncluded in every planAll plansGemini 3.8 Flash
Google · Text
Fast and cheap, for anything done in volume.
QualitySpeed1 cr per scriptAll plansMistral Large 4
Mistral AI · Text
Strong in European languages.
QualitySpeed2 cr per scriptAll plansDeepSeek V4
DeepSeek · Text
Lots of ideas for very few credits.
QualitySpeed1 cr per scriptAll plansLlama 4
Meta · Text
Captions, alt text and small text tasks.
QualitySpeed1 cr per scriptAll plansQwen 3.8
Alibaba · Text
Translates scripts, subtitles and articles.
QualitySpeed1 cr per scriptAll plansYour cloned voice
Mistral AI · Voice
Your videos narrated in your own voice.
QualitySpeed2 cr per minuteGrowth and upStudio voices
ElevenLabs · Voice
A library of native voices, ready to use.
QualitySpeed9 cr per minuteAll plansAvatar IV
HeyGen · Video
The AI creator who talks about your app, with lip-sync.
QualitySpeed100 cr per 15 s videoAll plansVeo 3.1
Google DeepMind · Video
The most realistic generated shots.
QualitySpeed390 cr per 8 s clipGrowth and upKling 3.0
Kuaishou · Video
Believable movement of people and objects.
QualitySpeed165 cr per 8 s clipGrowth and upSeedance 2.5
ByteDance · Video
Fast, cheap clips to dress up a video.
QualitySpeed145 cr per 8 s clipGrowth and upHailuo 3
MiniMax · Video
Expressive characters and lively scenes.
QualitySpeed125 cr per 8 s clipGrowth and upGemini 3 Pro Image
Google · Image
Realistic images and edits from your own screenshots.
QualitySpeed20 cr per imageAll plansSeedream 5
ByteDance · Image
Fast, stylized images.
QualitySpeed5 cr per imageAll plans
The scores
Quality and speed are scored out of 10 from our tests on app-marketing tasks. Use them to compare models within the same category.
Credits
Each generation uses credits depending on the model you pick, and the cost is shown before you start. Competitor analysis, scheduling and publishing are included in every plan.
Auto mode
If you'd rather not choose, Auto mode picks a model for each step based on the template or article type and the credits you want to spend.
Every model, side by side
| Model | Used for | Credits | Quality | Speed | Available on |
|---|---|---|---|---|---|
| Scripts & articles | |||||
| Claude Sonnet 5.5 Anthropic | Hook, Script, Articles | 4 per script | 9/10 | 7/10 | Growth and up |
| Claude Opus 5.5 Anthropic | Competitor audit, Content plan, Pillar pages | 8 per script | 10/10 | 4/10 | Growth and up |
| GPT-5.5 OpenAI | Ideas, Founder stories, Repurposing | 7 per script | 9/10 | 6/10 | Growth and up |
| Gemini 3.1 Pro Google | Video analysis, Page analysis, Your footage | Included | 9/10 | 7/10 | All plans |
| Gemini 3.8 Flash Google | Captions, SEO snippets, Variants | 1 per script | 7/10 | 10/10 | All plans |
| Mistral Large 4 Mistral AI | Scripts, Localization | 2 per script | 8/10 | 8/10 | All plans |
| DeepSeek V4 DeepSeek | Brainstorming, Hook variants | 1 per script | 7/10 | 8/10 | All plans |
| Llama 4 Meta | Captions, Checks | 1 per script | 7/10 | 9/10 | All plans |
| Qwen 3.8 Alibaba | Subtitles | 1 per script | 7/10 | 9/10 | All plans |
| Voiceover | |||||
| Your cloned voice Mistral AI | Voiceover, Your avatar, Translations | 2 per minute | 9/10 | 9/10 | Growth and up |
| Studio voices ElevenLabs | Voiceover | 9 per minute | 8/10 | 9/10 | All plans |
| Video clips | |||||
| Avatar IV HeyGen | AI UGC videos, Hook + demo, Your avatar | 100 per 15 s video | 9/10 | 7/10 | All plans |
| Veo 3.1 Google DeepMind | Lifestyle shots, Openers | 390 per 8 s clip | 10/10 | 3/10 | Growth and up |
| Kling 3.0 Kuaishou | Animated stills, B-roll | 165 per 8 s clip | 9/10 | 4/10 | Growth and up |
| Seedance 2.5 ByteDance | Transitions, Backgrounds | 145 per 8 s clip | 8/10 | 6/10 | Growth and up |
| Hailuo 3 MiniMax | Skits | 125 per 8 s clip | 8/10 | 5/10 | Growth and up |
| Images | |||||
| Gemini 3 Pro Image Google | Article covers, Mockups, Thumbnails | 20 per image | 9/10 | 7/10 | All plans |
| Seedream 5 ByteDance | Illustrations | 5 per image | 8/10 | 8/10 | All plans |
Questions about models
How are the quality and speed scores calculated?
They're our own scores out of 10, from testing each model on app-marketing tasks: writing a hook, a 30-second script, a 2,000-word article, a voiceover, a 5-second shot. They're meant to compare models with each other, not to judge them in general.
Why do some models cost more credits?
Credits follow what each model costs to run. A generated video clip takes far more compute than a script, hence the gap. The cost is always shown before you start a generation.
Do I have to pick a model every time?
No. Each template and article type comes with a default combination. You can change the model for any step, or turn on Auto mode, which picks based on the task and how many credits you want to spend.
Does the list of models change?
Yes. When a new model does better in our tests, we add it. When a model falls behind, we retire it, with notice if you use it in a saved template.
Are my videos, articles and voice used to train these models?
Your content is sent to the models only to produce your videos and articles. Where a provider offers it, we use terms that exclude training on your data.
Every model is included on the Growth plan: $199 a month, or $165 a month billed yearly.
See pricing