The Astra version seems to have used three.js r170, which is from October 2024. Sol used an even earlier version. GLM's code used the latest version, but I think it's just getting three.js@latest from jsdelivr so it's unlikely to be writing code against that version. Qwen on OpenCode also fetches from jsdelivr, but using a pinned version at r160.
I don't think any of these examples are using things like tone mapping so they're stuck in sRGB (AgX or ACES look much better), they're not using the node materials (good for programmatic texture implementation), and they're not doing anything cool like baking shadow environments or using post-processing effect.
They're nice, but I think they're showing how far behind AI models are on this sort of project rather than how good they are.
Really appreciate this perspective. It’s tempting to stop at being amazed at what these models can do, but my experience matches yours - they need help with details to do well.
Not tone mapping/shader graph is probably the least of the sins, these kinds of one-shots tend to produce hundreds of THREE.Geometry, massive matrix walks, etc. that explode once you move beyond a screen saver.
The models _can_ do it, but you need to ask for the right things. Most people don't, they'll usually blame the browser for being slow or ugly when they can't break through the THREE demo page wall.
A larger THREE.js project starts to look more and more like a game engine, so you pick and choose the parts you need. There's a ton of open source libs, most of the heavy components the big players use are open source, things like physics, mesh optimization.
Key AI-specific parts might be:
- a harness (so the agent can drive the thing)
- authoring pipeline (so you can bake/optimize assets)
- some sort of coherent renderer architecture (what are your assets, your passes, what's your shader graph).
Without some fundamentals here you are on the short road to falling off the cliff of tech debt and the AI will gladly drive you off of it until you ask for an expensive rewrite.
> would you create models independently
Yes. Pretty much any proven gamedev/asset pipeline is something frontier models are good at. Procedural systems, asset store, free content, Blender, Meshy.
Literally anything except "make a THREE.js scene" -- which is not a technique used in gamedev, beyond throwaway prototypes or demos. Which is what you will get if you ask for a THREE.js scene.
I wish there was another column with the estimated cost for each, with a specific date.
Ideally also finding somehow (not sure what would be the right away) what is publicly available before running the test. It's quite a different outcome if there are competitions, e.g. js13k, live code examples from books, even templates, on specific that topic. Visually here the results looks very very similar to the point that I can't help but wonder if it's the result from the short yet relatively descriptive prompt or because some template was always found and relied on.
I created my own in-browser terminal, so it can sit in chrome, accessible, and with some enhancements like animating the tab when the agent is doing some work. It's a layer on top of ttyd/tmux.
But I am annoyed at these GUIs implementing features I don’t care about. I want them to just wrap my harness and forward it to my iPhone, but they can’t help themselves from feature creep.
I did have some issues getting it installed on a headless server. I sort of gave up and installed the instance that I use as the remote server on a Debian + xfce machine I had laying around.
I found it appalling when working with remote setup — tabs disappearing and appearing after some time, screen hangs up, projects disappear. Tried to fix some issues with PR, but team is not accepting for 2w+.
I did try this - I found it overly dev focused - e.g. git +/- changes in left sidebar. I wanted something close to what I am use to with Codex/Claude - just for open models.
Nice. But these tests raise a question what results are reproducible, the final visual, time, tool calling or it's mostly noise. Like, have you tried same combination or model and harness multiple times?
It’s really interesting how much little choices make the result better or worse. Astra and one of the GLMs added bright lights, and thus looked so much better to my eye.
Genuinely happy with some of the Qwen 3.8 results (especially since I can run that model at Q8).
Interesting to see how much better (at this task) Pi (OMP) is over Opencode as a harness.
I’d love to see a few more with outcomes that are as easy to judge but less subjective.
I’ve got a toy project going to make a fun to watch battle simulator where an LLM (or two if playing vs) has to write programs that control multiple bots (each with their own line of sight and limited battle context) that have to coordinate and fight alongside each other. Goal is to have the LLM update the code based on current situations maybe 5-10 times in a 5 min simulated battle. Exploring even allow the bots to request new programming and score based on number of reprogram steps.
huh. goes to show how much browser selection impacts things. the OMP Qwen result is completely unresponsive in free cam on my iphone, whereas the opencode version is responsive, has multitouch zoom and multitouch pan. the glm models do really well with this as well, but all of the anthropic models have pretty gimped camera controls, even astra, which while pretty wont let me pan and only allows me to rotate about 120 degrees! I’m really impressed with the Qwen 27b OpenCode result.
I suppose the only thing that the prompt asks for is the cinematic view, and honestly they all kinda fail on the “subtle volumetric-style fog planes”, none of them have more fog when you get farther from a light source.
I love everything about OpenCode except for the stuff it outputs. On paper it has everything I want in a harness and more, but it tends to struggle to deliver the desired outcome. Last I used it the context was massive and tool calls were reliably unreliable. Is it worth revisiting as a daily driver?
I saw examples on X of people building amazing things with Three.js and models, and I figured out how to build them myself. But the results are mostly low-poly materials and crappy animations (pretty sure because I am not the best at these things). I thought it was my prompt (it was detailed), but it seems to be mostly a limitation of the models.
From all the examples I've seen, Astra does it really well, and I suspect it's because they wanted to attract game designers, so they trained the model more on 3D, animation libraries, etc.
Sol and Opus are pretty good with the right prompting and feedback framework.
Or, my standards are lower. It's sometimes hard to tell in these discussions whether people are talking about getting production quality results, or stuff that's good enough for a one off blog post.
I’m always confused, are all these shapes programmatically generated or are they downloaded from some source?
Also I think Astra looks the best and has the best functionality. Also shocked how much better GLM is on the Non Codex harnesses. Didn’t think it would make such a difference.
Would be nice if you could include cost in the table
Would be good to include Fable and Opus. Based on the results with Astra vs GLM (not even in the same ballpark), these are probably the realistic alternatives to consider.
Basically it can run these mini programs where each input might be another toolcall, so it can run without waiting for whole LLM response and ready the parameters async.
The Astra version seems to have used three.js r170, which is from October 2024. Sol used an even earlier version. GLM's code used the latest version, but I think it's just getting three.js@latest from jsdelivr so it's unlikely to be writing code against that version. Qwen on OpenCode also fetches from jsdelivr, but using a pinned version at r160.
I don't think any of these examples are using things like tone mapping so they're stuck in sRGB (AgX or ACES look much better), they're not using the node materials (good for programmatic texture implementation), and they're not doing anything cool like baking shadow environments or using post-processing effect.
They're nice, but I think they're showing how far behind AI models are on this sort of project rather than how good they are.
The fact that these tools are capable of doing anything at all is a miracle to me.
Really appreciate this perspective. It’s tempting to stop at being amazed at what these models can do, but my experience matches yours - they need help with details to do well.
Not tone mapping/shader graph is probably the least of the sins, these kinds of one-shots tend to produce hundreds of THREE.Geometry, massive matrix walks, etc. that explode once you move beyond a screen saver.
The models _can_ do it, but you need to ask for the right things. Most people don't, they'll usually blame the browser for being slow or ugly when they can't break through the THREE demo page wall.
Can you elaborate on this? how would you work on a larger THREE project? would you create models independently?
> how would you work on a larger THREE project?
A larger THREE.js project starts to look more and more like a game engine, so you pick and choose the parts you need. There's a ton of open source libs, most of the heavy components the big players use are open source, things like physics, mesh optimization.
Key AI-specific parts might be:
- a harness (so the agent can drive the thing)
- authoring pipeline (so you can bake/optimize assets)
- some sort of coherent renderer architecture (what are your assets, your passes, what's your shader graph).
Without some fundamentals here you are on the short road to falling off the cliff of tech debt and the AI will gladly drive you off of it until you ask for an expensive rewrite.
> would you create models independently
Yes. Pretty much any proven gamedev/asset pipeline is something frontier models are good at. Procedural systems, asset store, free content, Blender, Meshy.
Literally anything except "make a THREE.js scene" -- which is not a technique used in gamedev, beyond throwaway prototypes or demos. Which is what you will get if you ask for a THREE.js scene.
Anyone know of pertinent skill files to nudge the agents into these directions?
Don't use skills. Point it to a game engine or better yet full game repo or whatever and ask for what you want.
What’s a good way to drive models to use current versions?
I wish there was another column with the estimated cost for each, with a specific date.
Ideally also finding somehow (not sure what would be the right away) what is publicly available before running the test. It's quite a different outcome if there are competitions, e.g. js13k, live code examples from books, even templates, on specific that topic. Visually here the results looks very very similar to the point that I can't help but wonder if it's the result from the short yet relatively descriptive prompt or because some template was always found and relied on.
I've just pushed the transcripts of each if that helps. https://github.com/alvins82/hangar-harness-model-tests
Btw in my quest for a good desktop codex/claude like app - https://github.com/openchamber/openchamber - this seems to be front-runner. I pair it with OMP via https://github.com/alvins82/omp-openchamber-server/.
I wanted a powerful GUI+harness setup for open models so I could use/test as they came out.
I created my own in-browser terminal, so it can sit in chrome, accessible, and with some enhancements like animating the tab when the agent is doing some work. It's a layer on top of ttyd/tmux.
https://github.com/anilgulecha/ttydterm
I like Paseo https://github.com/getpaseo/paseo
But I am annoyed at these GUIs implementing features I don’t care about. I want them to just wrap my harness and forward it to my iPhone, but they can’t help themselves from feature creep.
I've been impressed with orca
https://github.com/stablyai/orca
I did have some issues getting it installed on a headless server. I sort of gave up and installed the instance that I use as the remote server on a Debian + xfce machine I had laying around.
I found it appalling when working with remote setup — tabs disappearing and appearing after some time, screen hangs up, projects disappear. Tried to fix some issues with PR, but team is not accepting for 2w+.
I did try this - I found it overly dev focused - e.g. git +/- changes in left sidebar. I wanted something close to what I am use to with Codex/Claude - just for open models.
i liked paseo too but its not optimized at all. let alone the feature bloat.
DeepSeek harness is for you
Nice. But these tests raise a question what results are reproducible, the final visual, time, tool calling or it's mostly noise. Like, have you tried same combination or model and harness multiple times?
I have on Qwen3.8 27B. Results are close enough across multiple runs.
So it's not reproducible.
It’s really interesting how much little choices make the result better or worse. Astra and one of the GLMs added bright lights, and thus looked so much better to my eye.
Genuinely happy with some of the Qwen 3.8 results (especially since I can run that model at Q8).
Interesting to see how much better (at this task) Pi (OMP) is over Opencode as a harness.
I’d love to see a few more with outcomes that are as easy to judge but less subjective.
I’ve got a toy project going to make a fun to watch battle simulator where an LLM (or two if playing vs) has to write programs that control multiple bots (each with their own line of sight and limited battle context) that have to coordinate and fight alongside each other. Goal is to have the LLM update the code based on current situations maybe 5-10 times in a 5 min simulated battle. Exploring even allow the bots to request new programming and score based on number of reprogram steps.
Interesting also that Astra and Sol didn't check visually using the browser. I suspect they would have done even better if they had.
huh. goes to show how much browser selection impacts things. the OMP Qwen result is completely unresponsive in free cam on my iphone, whereas the opencode version is responsive, has multitouch zoom and multitouch pan. the glm models do really well with this as well, but all of the anthropic models have pretty gimped camera controls, even astra, which while pretty wont let me pan and only allows me to rotate about 120 degrees! I’m really impressed with the Qwen 27b OpenCode result.
I suppose the only thing that the prompt asks for is the cinematic view, and honestly they all kinda fail on the “subtle volumetric-style fog planes”, none of them have more fog when you get farther from a light source.
Damn its unreal how much Astra is better than anything else :( Would be cool to see total cost
Should I be able to see the actual results? Because I don’t see a hangar anywhere.
File -> Open
If you don't have anything working check the console, maybe a WebGL issue.
Click Open in the file column
I love everything about OpenCode except for the stuff it outputs. On paper it has everything I want in a harness and more, but it tends to struggle to deliver the desired outcome. Last I used it the context was massive and tool calls were reliably unreliable. Is it worth revisiting as a daily driver?
i moved from opencode to omp and i don't think I'm ever going back
Why not ohmypi?
I saw examples on X of people building amazing things with Three.js and models, and I figured out how to build them myself. But the results are mostly low-poly materials and crappy animations (pretty sure because I am not the best at these things). I thought it was my prompt (it was detailed), but it seems to be mostly a limitation of the models.
From all the examples I've seen, Astra does it really well, and I suspect it's because they wanted to attract game designers, so they trained the model more on 3D, animation libraries, etc.
Sol and Opus are pretty good with the right prompting and feedback framework.
Or, my standards are lower. It's sometimes hard to tell in these discussions whether people are talking about getting production quality results, or stuff that's good enough for a one off blog post.
How do you judge which one produces the best result? They all look the same to me more or less.
I’m always confused, are all these shapes programmatically generated or are they downloaded from some source?
Also I think Astra looks the best and has the best functionality. Also shocked how much better GLM is on the Non Codex harnesses. Didn’t think it would make such a difference.
Would be nice if you could include cost in the table
What I'm missing in all of those one-shot tests:
How different are the results between multiple runs of the same setup?
I have done multiple runs on Qwen3.8 27B. Results are close enough across multiple runs qualitatively.
Would be good to include Fable and Opus. Based on the results with Astra vs GLM (not even in the same ballpark), these are probably the realistic alternatives to consider.
Am I missing something or does sorting by Duration not work? Chrome on Android mobile
cool, I've tested different harnesses with the same model. can you add jcode harness? I had great results with that https://grigio.github.io/ship-harness-bench/
Refreshing to see a post that gets straight to the point and not hide the data behind pages of prose.
Qwen with open code seemed like the best to me. It seemed like the best balance between performance and visuals
Astra has so much more detail. It did take 5x more time, but that's usually negligible when it replaces x hours of human output.
This is rad, the Sol one is probably the coolest.
So where did most of them get the idea of calling it hanger 07?
fun benchmark. did any harness consistently win regardless of model, or is it more about the pairing?
What is DSH / PTC?
Ptc is programmatic tool calling or code mode
Basically it can run these mini programs where each input might be another toolcall, so it can run without waiting for whole LLM response and ready the parameters async.
Deepseek Harness. PTC is one of the 'modes' it has.
to me i think the diffrenece is not that big?