GPT-5.6, Grok 4.5, Claude, and Muse Spark build the same 4 apps

Our last build-off hit the Hacker News front page, and the comments did not hold back. Fair enough, a lot of it was good feedback. So we took it, and with GPT-5.6 landing in three tiers (Sol, Terra, Luna) and Meta surprise-dropping a coding model (Muse Spark 1.1), we ran the whole thing again, bigger: twelve models, four apps, five attempts each.

What we changed based on your feedback:

• You wanted open-weights models in the mix. So we added GLM-5.2, Qwen 3.7 Plus, DeepSeek V4 Pro, and Kimi K2.6 as comparison points, all served via Fireworks.
• One attempt was weird, you said. Agreed. Every model now gets five attempts per task. Up top you get one sample run per model; each task table then says how many of the five we thought actually succeeded (and how we counted) and links the attempt we liked best; and every attempt is linked at the bottom so you can see how much these models swing run to run.
• "This isn't objective." Correct, and we are not pretending it is. We are not handing down a scientific verdict. We generated a big pile of artifacts, we are publishing all of them, and you can form your own opinion. Everything below is just our observations from watching the results.

Want to skip straight to poking at the raw builds? Jump to every attempt and run them yourself.

The lineup, twelve strong: the new GPT-5.6 Sol, GPT-5.6 Terra, and GPT-5.6 Luna; Meta's Muse Spark 1.1; Grok 4.5, GPT-5.5, Claude Opus 4.8, and Claude Fable 5; plus the open-weights crew: Qwen 3.7 Plus, DeepSeek V4 Pro, Kimi K2.6, and GLM-5.2.

Here is each task: our pick of the five attempts playing up top, the cost and time for all five just below, and links to every raw attempt at the bottom of the post so you can judge for yourself.

Task 1: Doom-style raycaster maze

First-person raycaster you walk with WASD, shaded walls with depth, floor and ceiling, collision.

Grok 4.5	Claude Opus 4.8	GPT-5.5	Qwen 3.7 Plus

Kimi K2.6	Claude Fable 5	DeepSeek V4 Pro	GLM-5.2

GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	Muse Spark 1.1

How we counted "playable": the only question we cared about was whether you could actually walk through the labyrinth, move and turn. If yes, it counted.

Model	Playable	Best build	Cost (5 runs)	Avg time	Our take
Grok 4.5	5/5	#4	$0.27	62s	Great for the price.
Claude Opus 4.8	4/5	#2	$0.54	48s	Consistent but dry.
GPT-5.5	4/5	#1	$1.44	138s	Best overall before GPT-5.6.
Qwen 3.7 Plus	2/5	#4	$0.13	43s	—
Kimi K2.6	2/5	#2	$1.37	88s	—
Claude Fable 5	3/5	#1	$2.35	107s	Good results, but less consistent.
DeepSeek V4 Pro	3/5	#4	$0.30	318s	—
GLM-5.2	0/5	—	$0.12	133s	Rendered good detail, but the character could not move in any attempt.
GPT-5.6 Sol	5/5	#5	$1.35	120s	Best results: consistent, better than GPT-5.5, and with more detail in the game overall.
GPT-5.6 Terra	3/5	#1	$0.44	39s	Good detail, but less consistent, I couldn't always walk.
GPT-5.6 Luna	5/5	#5	$0.15	23s	Great results, though not as good as GPT-5.5 in my opinion.
Muse Spark 1.1	2/5	#1	$0.55	169s	Surprisingly awesome when it worked. Three of five were broken, but the working ones were on par with Fable and Sol, and better than the Grok and Opus ones.

Overall, Claude did worse than we expected here. GPT outperformed every other model, Grok was a genuinely usable alternative at its price point, and Muse Spark was a real surprise on the runs that actually worked.

Task 2: 3D Rubik's Cube (scramble + solve)

Build a colorful, 3D-looking Rubik's Cube with Scramble and Solve buttons that visibly animate the rotations.

Grok 4.5	Claude Opus 4.8	GPT-5.5	Qwen 3.7 Plus

Kimi K2.6	Claude Fable 5	DeepSeek V4 Pro	GLM-5.2

GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	Muse Spark 1.1

How we counted a "clean solve": we scrambled and solved the cube and only counted an attempt if both animations ran smooth, no glitches, no color changes.

Model	Clean solve	Best build	Cost (5 runs)	Avg time	Our take
Grok 4.5	3/5	#4	$0.65	191s	Good, simple results.
Claude Opus 4.8	0/5	#1	$0.56	44s	No perfect example, everything had a small issue and the cube's colors would change, though some were close.
GPT-5.5	4/5	#4	$1.36	136s	Good results, with small nits like flickering colors or not-quite-smooth rotation animations.
Qwen 3.7 Plus	1/5	#5	$0.07	24s	When it worked, it looked great!
Kimi K2.6	1/5	#4	$1.05	59s	Not great.
Claude Fable 5	5/5	#2	$2.03	92s	No notes, really nice results.
DeepSeek V4 Pro	1/5	#3	$0.35	380s	—
GLM-5.2	0/5	—	$0.08	89s	Surprisingly, no working results.
GPT-5.6 Sol	4/5	#5	$1.06	72s	Great results. One had weird animations and one rendered an all-black cube (interesting choice), but it did great on the ones that worked.
GPT-5.6 Terra	4/5	#4	$0.42	38s	Fine, with weird scramble animations, felt like a small step up from GPT-5.5.
GPT-5.6 Luna	0/5	—	$0.14	24s	Usually rendered correctly at first, but scrambling immediately broke it.
Muse Spark 1.1	2/5	#1	$0.54	182s	A step above the OSS models, but not something I'd use over Grok for the price.

Overall, we were surprised GPT underperformed here given its clear 3D lead in the raycaster. Claude did an amazing job again, though it was Fable carrying it with a clean five-for-five, while Opus, oddly, couldn't land a single flawless solve.

Play all five of any model live: Grok · Opus · Qwen (swap the number 1-5 for other attempts).

Task 3: Calculator

Digits, operators, clear, equals, correct operator precedence, real calculator look.

Grok 4.5	Claude Opus 4.8	GPT-5.5	Qwen 3.7 Plus

Kimi K2.6	Claude Fable 5	DeepSeek V4 Pro	GLM-5.2

GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	Muse Spark 1.1

How we counted "working": nothing exhaustive, just basic calculations like (((5 × 5) − 100) / 10) to see how each one handled order of operations and rendered the result.

Model	Working	Best build	Cost (5 runs)	Avg time	Our take
Grok 4.5	5/5	#5	$0.37	110s	Simple, consistent results, no extra styling.
Claude Opus 4.8	5/5	#1	$0.46	35s	—
GPT-5.5	4/5	#1	$0.91	75s	Sometimes added extra buttons, and it tried to render in 3D, which often got cut off.
Qwen 3.7 Plus	4/5	#4	$0.04	12s	Worked and pretty consistent, though one couldn't handle negative numbers.
Kimi K2.6	0/5	#5	$0.76	50s	Worked, but couldn't handle negative numbers.
Claude Fable 5	5/5	#1	$1.22	48s	My personal favorite, purely on style.
DeepSeek V4 Pro	3/5	#3	$0.28	342s	One had out-of-order numbers, and one result didn't render.
GLM-5.2	2/5	#1	$0.07	63s	Great results when it worked.
GPT-5.6 Sol	5/5	#1	$0.84	61s	Tries to go overboard on styling like GPT-5.5. That helped in the raycaster, but here it was annoying, some styling was off instead of clean and consistent.
GPT-5.6 Terra	4/5	#1	$0.28	20s	Good results when it worked.
GPT-5.6 Luna	5/5	#1	$0.11	16s	Good, consistent results, felt similar to Grok's.
Muse Spark 1.1	5/5	#5	$0.33	94s	Pretty good overall with no glaring issues, on par with Grok 4.5. Occasionally out-of-order or oddly-placed buttons cost it some style points.

Overall, this was clearly Claude's best work: both Opus and Fable nailed all five, and Fable's was our favorite on style. GPT-5.6 Sol tries to go overboard with styles and render the calculator in 3D similar to Fable. However, it doesn't nail the styles, which leads to a worse overall experience. The simpler GPT models seemed to do a better job just because the experience worked out of the box. GLM-5.2, reasoning-off, also came back from nine-minute failures to a snappy, fraction-of-a-cent build.

Task 4: Conway's Game of Life

Grid canvas, Play/Pause/Step/Randomize/Clear, click to toggle cells, animated generations.

Grok 4.5	Claude Opus 4.8	GPT-5.5	Qwen 3.7 Plus

Kimi K2.6	Claude Fable 5	DeepSeek V4 Pro	GLM-5.2

GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	Muse Spark 1.1

We did not run a separate five-attempt scoring pass on Life, so here it is just cost and time plus the general impression below.

Model	Cost (5 runs)	Avg time
Grok 4.5	$0.14	38s
Claude Opus 4.8	$0.48	35s
GPT-5.5	$1.06	79s
Qwen 3.7 Plus	$0.04	11s
Kimi K2.6	$0.93	53s
Claude Fable 5	$1.27	48s
DeepSeek V4 Pro	$0.25	304s
GLM-5.2	$0.10	121s
GPT-5.6 Sol	$0.99	62s
GPT-5.6 Terra	$0.36	25s
GPT-5.6 Luna	$0.13	18s
Muse Spark 1.1	$0.32	98s

Grok 4.5 did well here, but the bigger takeaway is that this task was simple enough for the OSS models to do extremely well on. There is probably enough open example code for Game of Life out there that they were able to do a much better job at a way lower cost. Qwen 3.7 Plus and GLM-5.2 are clearly the go-to for something like this, but I would not rely on them generally, the other tasks show they still struggle with genuinely novel or more complex work.

The receipts: raw speed and cost (short answers)

Separate question, separate table. This is our standard latency harness (three short prompts, five reps, 400-token cap), not the build tasks. tok/s is output tokens over wall-clock, uniform for all.

Model	Median latency	Throughput	Cost / reply
GPT-5.6 Luna	1.0s	97 tok/s	$0.001
GPT-5.6 Terra	1.5s	62 tok/s	$0.001
GPT-5.6 Sol	1.8s	45 tok/s	$0.003
Qwen 3.7 Plus	2.1s	204 tok/s	$0.001
Claude Opus 4.8	2.5s	44 tok/s	$0.004
GPT-5.5	3.0s	45 tok/s	$0.003
Grok 4.5	3.0s	112 tok/s	$0.003
Muse Spark 1.1	3.1s	125 tok/s	$0.002
Kimi K2.6	4.5s	83 tok/s	$0.01
Claude Fable 5	6.6s	30 tok/s	$0.01
GLM-5.2	7.0s	58 tok/s	$0.001
DeepSeek V4 Pro	9.3s	37 tok/s	$0.001

One honest caveat: several open-weights buffered their whole reply in a burst and hit the 400-token cap, so their tok/s is a ceiling, not a true decode rate. The clear read: the GPT-5.6 tiers are the snappiest models here on short prompts (Luna answers in about a second), Qwen is absurdly cheap and fast, and DeepSeek and GLM are the slowpokes, which matched the agony of generating their apps.

Bonus: draw a horse riding an astronaut (SVG, best of 5)

One-shot SVG, no libraries. We show each model's best-of-five (we prefer a strictly-valid SVG, then the most detailed).

Grok 4.5	Claude Opus 4.8	GPT-5.5	Qwen 3.7 Plus

Kimi K2.6	Claude Fable 5	DeepSeek V4 Pro	GLM-5.2

GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	Muse Spark 1.1

Personally, Claude Fable does a great job with the SVG rendering. It's funny most of the time as well and came up with good quality results. The GPT-5.6 models were surprisingly lackluster here since none of them included clean renderings of the horse or the astronaut. Grok 4.5 did pretty well here as well.

Bonus 2: Elon and Bezos watch a Blue Origin landing (SVG, best of 5)

A harder scene by request: two recognizable tech-billionaire caricatures watching a Blue Origin booster land on a pad in the open ocean. This one leans on composition and likeness. Again, Claude Fable sweeps the field: it produces great detail with a clean render, a shiny spot on Bezos' forehead, and smoke around the landing pad. GPT again produces pretty cartoony results; if that's something you prefer then maybe you can try improving the one-shot results it produces, since there are always some minor errors with the generations. The OSS models, GLM-5.2 and Qwen 3.7, also did really well for this task.

Grok 4.5	Claude Opus 4.8	GPT-5.5	Qwen 3.7 Plus

Kimi K2.6	Claude Fable 5	DeepSeek V4 Pro	GLM-5.2

GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	Muse Spark 1.1

TL;DR

• The frontier still wins the hard tasks, and it is not particularly close on the complex ones. GPT-5.6 Sol and Claude Fable 5 were the standouts: Sol excelled at the raycaster, Fable at the Rubik's cube. Everything else did decently, but the best results are still at the frontier.
• There is a clear gap between SOTA and open-weights on genuinely novel or complex work, and you can see it in the raycaster and cube. But on a simple, well-trodden task like Game of Life, the OSS models hold their own, there is enough example code out there that Qwen 3.7 and GLM-5.2 nail it at a fraction of the cost. Use them for that class of problem; just do not lean on them generally, because the other tasks show they still struggle with the hard stuff.
• Grok 4.5 is genuinely "Opus level" on some tasks. If I cared about cost for my business, I would reach for it as a secondary execution model without hesitation. See Grok 4.5 vs Opus 4.8 and Grok 4.5 vs GPT-5.5.
• Muse Spark 1.1 pleasantly surprised me. I was not really expecting much, and it felt a step below Grok 4.5 but generally better than the open-weights. It is a real debut, though not something I would reach for just yet.

The takeaway holds even after launch day: the newest, most expensive flagship is not the automatic winner. Want to run these prompts yourself? Every model here is on one TryAI account, pay-as-you-go. Browse the models and start your own build-off.

Play with every attempt

Every raw build we generated, twelve models times five attempts per task. Click any number to open that exact attempt and poke at it yourself; the swings between attempts are the point.

Raycaster maze

Model	Attempts
Grok 4.5	1 · 2 · 3 · 4 · 5
GPT-5.6 Sol	1 · 2 · 3 · 4 · 5
GPT-5.6 Terra	1 · 2 · 3 · 4 · 5
GPT-5.6 Luna	1 · 2 · 3 · 4 · 5
Muse Spark 1.1	1 · 2 · 3 · 4 · 5
GPT-5.5	1 · 2 · 3 · 4 · 5
Claude Opus 4.8	1 · 2 · 3 · 4 · 5
Claude Fable 5	1 · 2 · 3 · 4 · 5
Qwen 3.7 Plus	1 · 2 · 3 · 4 · 5
DeepSeek V4 Pro	1 · 2 · 3 · 4 · 5
Kimi K2.6	1 · 2 · 3 · 4 · 5
GLM-5.2	1 · 2 · 3 · 4 · 5

Rubik's Cube

Model	Attempts
Grok 4.5	1 · 2 · 3 · 4 · 5
GPT-5.6 Sol	1 · 2 · 3 · 4 · 5
GPT-5.6 Terra	1 · 2 · 3 · 4 · 5
GPT-5.6 Luna	1 · 2 · 3 · 4 · 5
Muse Spark 1.1	1 · 2 · 3 · 4 · 5
GPT-5.5	1 · 2 · 3 · 4 · 5
Claude Opus 4.8	1 · 2 · 3 · 4 · 5
Claude Fable 5	1 · 2 · 3 · 4 · 5
Qwen 3.7 Plus	1 · 2 · 3 · 4 · 5
DeepSeek V4 Pro	1 · 2 · 3 · 4 · 5
Kimi K2.6	1 · 2 · 3 · 4 · 5
GLM-5.2	1 · 2 · 3 · 4 · 5

Calculator

Model	Attempts
Grok 4.5	1 · 2 · 3 · 4 · 5
GPT-5.6 Sol	1 · 2 · 3 · 4 · 5
GPT-5.6 Terra	1 · 2 · 3 · 4 · 5
GPT-5.6 Luna	1 · 2 · 3 · 4 · 5
Muse Spark 1.1	1 · 2 · 3 · 4 · 5
GPT-5.5	1 · 2 · 3 · 4 · 5
Claude Opus 4.8	1 · 2 · 3 · 4 · 5
Claude Fable 5	1 · 2 · 3 · 4 · 5
Qwen 3.7 Plus	1 · 2 · 3 · 4 · 5
DeepSeek V4 Pro	1 · 2 · 3 · 4 · 5
Kimi K2.6	1 · 2 · 3 · 4 · 5
GLM-5.2	1 · 2 · 3 · 4 · 5

Game of Life

Model	Attempts
Grok 4.5	1 · 2 · 3 · 4 · 5
GPT-5.6 Sol	1 · 2 · 3 · 4 · 5
GPT-5.6 Terra	1 · 2 · 3 · 4 · 5
GPT-5.6 Luna	1 · 2 · 3 · 4 · 5
Muse Spark 1.1	1 · 2 · 3 · 4 · 5
GPT-5.5	1 · 2 · 3 · 4 · 5
Claude Opus 4.8	1 · 2 · 3 · 4 · 5
Claude Fable 5	1 · 2 · 3 · 4 · 5
Qwen 3.7 Plus	1 · 2 · 3 · 4 · 5
DeepSeek V4 Pro	1 · 2 · 3 · 4 · 5
Kimi K2.6	1 · 2 · 3 · 4 · 5
GLM-5.2	1 · 2 · 3 · 4 · 5