Vibedex · The AI Filmmaking Report · Issue 02 · Platform Rankings, part 2
AI Filmmakers’ Choice
AI Filmmaking Platform Rankings 2026, Part 2
Seventeen more working filmmakers, from a feature-film VFX compositor to the director of a feature made entirely with generative AI, vote for the three platforms they open on deadline, and say why.
Part 1 asked twenty filmmakers one blunt question: it's the night before a deadline, which platform do you open? Part 2 is the next seventeen ballots, in the order they arrived, from the same survey. Same question, same scoring, and the same follow-up: why that order?
Higgsfield won again. Four first-place votes and a place on nine of the seventeen ballots, with Magnific second and OpenArt third. What's new is how differently this group works. Several of them don't treat a platform as the place the film gets made. They treat it as one tool inside a pipeline they built themselves.
VibeDex is an independent AI comparison engine. We blind-benchmark the models ourselves and map what creative platforms can actually do. This report is the human half: what working filmmakers choose when it's their own time and their own credits on the line.
The short answer
In part 2 the panel's #1 is Higgsfield (21 points, 4 first-place votes), then Magnific (13) and OpenArt (10). Take the partner votes out and Higgsfield and Magnific do not move at all. OpenArt drops to 4.
Why there are two columns
7 of the 17 have a creator, partner or ambassador relationship with at least one platform. So we count every vote, then count again with each person's partner platforms taken out. A platform that holds up in the second column is one filmmakers pick even without a relationship.
01
Which platform do AI filmmakers rank first?
Higgsfield, again. The scoring is the same as part 1: 3 points for a #1, 2 for a #2, 1 for a #3. Higgsfield took 21 points from nine ballots, and four of those nine put it first. It was on more than half the ballots in this part, which no other platform managed.
17 ballots, 3/2/1 points. Platforms on fewer than five ballots show total points only.
The tail is long again. Twenty different platforms and tools turn up across seventeen ballots, and only Higgsfield, Magnific and OpenArt appear on five or more. One clear leader and a scattered field behind it, the same shape as part 1.
02
Does the ranking survive without partner votes?
Yes, and at the top it makes no difference at all. Creator programs are now how filmmakers get credits, early access and an audience, so we asked everyone to name their ties, then ran the count again without those votes.
7 of 17disclosed at least one partner, creator or ambassador relationship
11 of 17voted for no platform they partner with
0partner votes for Higgsfield or Magnific, the top two
Higgsfield and Magnific keep every point. None of the filmmakers who ranked them has a partner relationship with them, so the second column is identical to the first. OpenArt is the one that moves: it falls from 10 points to 4, because three of the five ballots that ranked it came from members of its own creator program.
That's a different picture from part 1, where half the panel sat in Dreamina's creator program. Here two filmmakers name it. If you have no deal with anyone, the second column is the one to read, and in this part it tells the same story as the first.
Platforms are good for early testing and sometimes for getting inspirations for workflow building.
03
Why do they pick what they pick?
The explanations rarely mention a model. They talk about what a platform costs, how fast it renders and how much room it leaves you. Four reasons came up.
What it costs
Cost came up first. One filmmaker took out a subscription because it was the cheapest and the best value, built projects on it, and pitched clients with the work. Another keeps Magnific as an everyday aggregator because it was the best on usability and pricing, and another because it now offers free video models. A third picked Runway for one reason: unlimited Seedance generations.
How fast it renders
Speed is what put Dreamina first on one ballot. The filmmaker says it beats third-party platforms running Seedance 2.5 through an API, and that's why it sits first on their ballot.
How much control you get
Two filmmakers put Figma Weave first, for different reasons. One came from compositing, where node-based tools are second nature. The other says it combines models and open-source workflows in ways a classic aggregator doesn't, with more freedom than Magnific and none of ComfyUI's complexity.
How much is under one roof
Higgsfield's supporters talk about range and direction. One uses it for the first character images before moving to Runway for video and post. Another says it fits how a director thinks: framing, lighting, camera movement and atmosphere. A third uses it for its MCP tools. invideo, picked first by one filmmaker and also ranked second by another, wins on something different: agents that work together on a project.
On models, the same two names carry most of the clips. Five of the seventeen made the film they shared with Kling, and three with Seedance, reached through different front doors, from OpenArt to Higgsfield to Runway to Magnific. When everyone has the same engines, the platform choice is about the car.
04
What still costs AI filmmakers the most time?
Control, again, and then money. Every contributor told us what platforms need most, and four themes cover most of the answers.
Consistency across a scene (5 of 17). Characters, costumes and places that hold from one shot to the next. Brian Lerch, who makes serialized drama, wants a model that remembers the geometry of a set: once a wide shot establishes where things are, he wants to cut to a close-up or a reverse and have everything still be where it was.
Fixing one thing without starting over (5 of 17). A hand, an expression, a mask, one costume. Vitaly Mostovoy wants to edit specific elements of a finished video without regenerating the scene. Priya Ayengar, coming from VFX, goes further and asks for pixel-perfect masks and real control over lighting.
Cost and billing (5 of 17). Priya says it isn't cost-effective once you're iterating. Marcelo Presotto asked for a watermarked, low-resolution preview, with credits charged only once a shot is approved. John Darbonne wants a billing statement before he's charged.
Camera and blocking (4 of 17). Tom Kinnersly wants advanced camera moves and character blocking without leaving the platform for Blender. Marta Solik asks for precise camera and performance control, and a timeline that holds continuity across a whole sequence.
One filmmaker answered a different question. Professional work, they say, can't be mastered on platforms: platforms are toys and ComfyUI-style systems are tools, and their feature film pipeline runs on direct API calls, deterministic scripts and ComfyUI workflows.
Getting one beautiful shot is only the beginning.
05
What the panel would tell you
Story comes first again. Katya Grechko put it in seven words: make a story, not just a picture. Christina J gave the order: story first, references second, prompts third. The rest of the advice was practical.
Plan before you generate
Write a shot list and a storyboard first, so every generation has a purpose and you don't waste credits.
Build references first
Character, prop and location sheets before any video. Samo Rera approves still frames first and builds the movement around them.
Test cheap
Use low resolution to test a shot, and use a draft mode when the platform has one.
Avoid the default look
Switch image models, or use film references, to escape the generic one.
Mix the models
Different models are good at different things. Test them, combine them and start editing early.
Think like a filmmaker
Camera placement, coverage and eyelines. Sometimes the shot that matters is the person listening. And check the cut: Tom Kinnersly warns that the best video models will happily break the 180-degree rule, so get a second opinion on your edit before you publish.
A few went further and told us to build your own pipeline. Don't live in a platform, wrap your own workflow around the pieces, and expect to fail ten times before the first one works. And one filmmaker kept it to two words: minimal prompts.
06
What comes next
This is the second and last part of Issue 2. Together, the two parts cover 37 filmmakers. We'll publish the combined ranking of both parts next.
Every ballot from now on also rates how well each contributor's top platforms keep a character and a look consistent across shots, and what one finished, usable clip costs once the retries are counted. These are the two things this panel complained about most. Part 2's ballots came in before that change, so the ratings start with the next issue.
The Panel
The Contributors
All seventeen contributors are credited by name below, and every answer is counted in every table. Full bios and their featured work are in the Appendix.
Emmy award winning editor and AI filmmaker as well as edit consultant, panellist and juror for the recent Big Screen Hack at IMAX London.
ALL the current best video models will quite happily break the '180 degree rule' and ruin your shot flow so always get a second opinion on your edit and story logic before you publish.
Makes films about geopolitical scenarios and world-order problems.
Keep making projects, thinking, ideation, brainstorming, is the skill.
Methodology
Contributors are working AI filmmakers recruited individually through their published work. Each completed the same written survey between 15 and 26 October 2026: their #1, #2 and #3 creative platforms from the past three months, why that order, what platforms most need to improve, one tip for other filmmakers, and any partner, creator or ambassador relationships. This part covers the other 17 ballots that met three conditions: an explanation of the ranking at least 50 characters long, one ballot per person, and no vote for a platform the contributor founded. Ballots that arrived after this part was closed are not in it.
Scoring gives 3 points for a #1, 2 for a #2 and 1 for a #3. The "without partner votes" column removes, for each contributor, any vote for a platform they named a relationship with. Counts of themes in sections 03 to 05 are the number of contributors who raised each point unprompted. Vibedex took no payment from any platform named in this report, and no platform saw the results before publication.
Appendix
Contributor bios and featured work
Full biographies, and the piece each contributor chose to share, with the tools they used to make it.
Tom Kinnersly
Editor and AI filmmaker
Tom Kinnersly is an Emmy Award winning editor with 20 years experience across drama, documentary, news and advertising. Clients include the BBC, HBO, CNN, ITV, Disney+, Sky and Working Title as well as major brands. Working with GenAI video tools since 2024, he is excited to see AI filmmaking evolve this year into the production-ready system it deserves to be. He sees professional editing as the vital component in AI filmmaking teams that it has always been in traditional live action filmmaking.
Priya Ayengar
Senior VFX compositor
Senior VFX compositor with 12+ years in feature film, having worked at ILM, MPC, Scanline and Luma on productions including Ahsoka, Thunderbolts, Deadpool and Project Hail Mary.
Made with Figma Weave with Pixverse 6, Google Flow with Nano Banana 2.1
Marcelo Presotto (Mr. Press)
Filmmaker, director and AI director
Brazilian filmmaker, director, screenwriter and AI director, and one of the pioneers of AI filmmaking in Brazil. Four decades as an editor and post-producer, more than 300 commercials and openings for TV Globo, and the feature film 2, released on Prime Video. In 2026 he directed Super Me, produced entirely with generative AI from an original screenplay written by humans, which premiered on 3 October 2026 at the Eden Théâtre in La Ciotat, France. Its trailer won Best National Trailer and Best Sound Design at the All That Moves International Film Festival.
Egyptian filmmaker, creative director and founder of ElCreators, combining cinematic storytelling with AI production. His short film The Seven Books won five international awards within 30 days and screened at IMAX in London.
Filmmaker, editor and producer with more than 20 years in television and documentary production, and founder of The Daily Then and Big Little Cinema. His current work spans historical reconstruction, documentary and serialized vertical drama using AI-assisted filmmaking.
Made with Magnific, Kling, Suno, ElevenLabs, Premiere Pro, Photoshop, ChatGPT
John Darbonne
Filmmaker and professor of media production
Assistant Professor of Media Production at Texas A&M University–Corpus Christi, with 25 years of Hollywood experience and 32 feature credits, including Automatic (SXSW) and his award-winning feature Companion, which he wrote, directed and produced and which sold to Level 33 Entertainment. In Spring 2024 he taught the first for-credit AI Filmmaking course in Texas, and he developed PedaGPU, a curriculum that puts professional AI filmmaking tools in the hands of students at regional universities.
Made with Live actors, greybox, Nano Banana Pro Edit, Fal.ai, Beeble, Magnific, Topaz, ComfyUI
Karim Gallimore
AI filmmaker
AI filmmaker and founder of From Sketch to Soul, blending a passion to create with an engineering background. A three-time Mindvalley AI Artist Award winner, he treats his image, video, voice and music tools like instruments in an orchestra, so they serve impactful stories instead of showing off the software.
Made with First Concepts, Higgsfield (Kling), GPT Image 2.0, ElevenLabs, Premiere Pro, ChatGPT, Claude
Katya Grechko
Creative director and AI visual storyteller
Creative director and AI visual storyteller. Award winner at the Tirana AI Film Festival and finalist at the Africa AI Awards, working in music videos and digital art.
Made with Kling, Runway, Midjourney, Canva, Photoshop, Final Cut Pro
Tobias Haucke
Filmmaker building AI pipelines
Filmmaker and feature film producer from Cologne, with a background spanning documentaries, commercials and animation. Now building AI pipelines for feature film production and co-creating AI-driven films at AIBoys Studio. Host of the YouTube channel “Prompt · Try · Repiet”, testing the latest tools with one belief: AI doesn’t replace good filmmaking, it opens up more room to experiment.
Marta Solik
AI creative director
Creates with AI across film, image, fashion, branding and digital experiences, focused on turning emerging tools into real creative workflows.
Made with Cascade Studio, GPT Image 2, Seedance 2.0, ElevenLabs, Premiere Pro
Vitaly Mostovoy
AI filmmaker and generative visual artist
Vitaly Mostovoy is an AI filmmaker, generative visual artist and creative technologist specializing in cinematic storytelling, AI-driven film production and visual effects. His work combines traditional filmmaking principles with emerging generative technologies, focusing on visual consistency, atmosphere and narrative-driven storytelling.
Deputy Festival Director of the Naija AI Film Festival, bringing creative leadership and operational expertise to the intersection of technology and storytelling. Passionate about AI governance and content distribution, she is dedicated to shaping sustainable frameworks and expanding global markets for African creators.
Creative director specializing in high-quality AI films, brand visuals and content for AI companies, bridging traditional storytelling with the newest generative tools.
AI-hybrid filmmaker and generative AI prompt engineer
Nigerian AI-hybrid filmmaker, generative AI prompt engineer and tech entrepreneur, specializing in human-AI systems thinking and the responsible integration of artificial intelligence into creative industries.
Part 1 has the first twenty ballots, and Issue 1 covers the models the panel ships with. The AI Ads and UGC Platform Rankings cover the studios making ad creative for brands.