AI/ML API

Platform for developers and entrepreneurs looking to integrate cutting-edge AI capabilities into their products. Team: @skinbagwbones @schwarzerrttr

[AI/ML API @aimlapi]

We've tested the Qwen 3.8 Max preview and it absolutely rocks!

Each scene is built in one shot, a single Three.js file with nothing else added:
• Odysseus' war galley riding the waves
• Trojan horse
• Greek war helmet

Qwen managed to execute the task faster than the other models, while output quality is relatively the same across all of them. If the pricing stays close to Alibaba's previous flagships, we'll get one of the best speed/quality ratios on the market — perfect for running agents at scale.

Soon live on AI/ML API.

[Qwen @Alibaba_Qwen]

Qwen3.8 is launching and going open-weight soon!🌐

With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5.

You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba’s Token Plan, Qoder, and QoderWork. Be among the very first to try it out.

Can't wait to hear what you build. Stay tuned! 🚀 

Token Plan
international：https://t.co/YRvcGdB9Bv
China：https://t.co/PKMUNwUuRp

[AI/ML API @aimlapi]

@BasileusBas Gemini is awesome. We've already made several runs and it is on 🔥

[AI/ML API @aimlapi]

We've tested the Qwen 3.8 Max preview and it absolutely rocks!

Each scene is built in one shot, a single Three.js file with nothing else added:
• Odysseus' war galley riding the waves
• Trojan horse
• Greek war helmet

Qwen managed to execute the task faster than the other models, while output quality is relatively the same across all of them. If the pricing stays close to Alibaba's previous flagships, we'll get one of the best speed/quality ratios on the market — perfect for running agents at scale.

Soon live on AI/ML API.

[Qwen @Alibaba_Qwen]

Qwen3.8 is launching and going open-weight soon!🌐

With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5.

You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba’s Token Plan, Qoder, and QoderWork. Be among the very first to try it out.

Can't wait to hear what you build. Stay tuned! 🚀 

Token Plan
international：https://t.co/YRvcGdB9Bv
China：https://t.co/PKMUNwUuRp

[thehype. @thehypedotnews]

gemini 3.6 flash vs qwen3-max vs gpt 5.6 sol vs kimi k3

@OfficialLoganK and @GoogleAIStudio recently shipped gemini 3.6 flash – their new workhorse. key facts:

• 17% fewer output tokens than 3.5 flash on the artificial analysis index, up to 65% on deepswe, at a lower price – $1.50/1m in, $7.50/1m out

• step up in coding, knowledge work and computer use (osworld-verified 83.0 vs 78.4); fewer reasoning steps and tool calls per task

• pitched as the efficiency sweet spot for scaling agents, not the frontier

the other side of the ring: @alibaba_qwen's qwen3.8-max-preview, unveiled july 19 at waic. 2.4t params, and in alibaba's own release the team calls it "compatible with leading frontier models, second only to fable 5"

one bets on speed and price, the other on raw capability. so we gave both the same hard job – rebuild a real building in 3d from a single photo

every model run through @aimlapi, our partner platform – one api, all the frontier labs in one place

our test – 3 prompts, single-file html, procedural @threejs via cdn, no assets. the model gets one photo and a spec, and has to return an explorable scene: massing, materials, glazing, interior, orbit, day/dusk:

1. ligar bay bach – nz cantilevered cliffside house, blonde timber + black steel + glass

2. blackwood ridge (the sleeve house) – charred-timber wedge with a raking prow

3. fallingwater – frank lloyd wright's concrete trays over a waterfall

results (summed across all 3 houses):

- cost
#1 qwen3-max – $0 (free preview)
#2 gemini 3.6 flash – $0.61
#3 kimi k3 – $1.07
#4 gpt 5.6 sol – $1.75

- tokens
#1 gpt 5.6 sol – 53,889
# qwen3-max – 65,875
#2 gemini 3.6 flash – 69,486
#4 kimi k3 – 170,803

- lines of code
#1 qwen3-max – 2,753
#2 kimi k3 – 2,793
#3 gpt 5.6 sol – 3,395
#4 gemini 3.6 flash – 3,370

overall: no single winner – kimi builds the best-looking scene, qwen the most detailed one for free and on the fewest lines, gpt the highest fidelity but at ~3x the price, and gemini flash the cheapest that still holds up

so we read the code

qwen3-max – the tidiest file of the four. zero external assets: procedural canvas textures for every surface and hand-written glsl shaders for the waterfall + pool. bakes its own pmrem env map in-scene, lerps day/dusk per-frame, and instances the forest with rejection sampling that carves a clearing around the house. the most economical programmer here – fewest lines, most packed into them

gemini 3.6 flash – leaner, but real three.js underneath: correct cantilevers, red mullions, working massing. the headline is that a flash-tier model writes this at all. 3.5 flash gave you threadbare boxes; 3.6 gives you a building. thinner material and lighting layer than qwen, but no longer reads as a budget model

our take:

• kimi k3 is the top designer of the group – better eye than qwen3.8, best composition and lighting on load

• qwen3.8 is the surprise: error-free, fast, plenty of detail – and it does it on the fewest lines of code. efficient and clean at the same time

• gemini 3.6 flash is the story for the labs – flash models are no longer a cheap tier you settle for. they're starting to compete with the full-size models outright

follow @thehypedotnews for 24/7 ai news, analysis and breakdowns

[AI/ML API @aimlapi]

@lablabai @FeatherlessAI lablab 🩷

[AI/ML API @aimlapi]

@saen_dev we tweaked the prompt a lot to make results consistent in one shot

[AI/ML API @aimlapi]

@AvoraMoney kinda funny 🙈

[AI/ML API @aimlapi]

@mktpavlenko @testingcatalog The scenes were built auto-rotating

[AI/ML API @aimlapi]

@HarakiriInu qwen 3.7 was not a joke too :)

[AI/ML API @aimlapi]

@A_A_Toulkeridis yup, wavy

[AI/ML API @aimlapi]

@snskritinaruka We've also tested it on long horizon tasks - and it performed great too!

[AI/ML API @aimlapi]

@atomicagent_io @schwarzerrttr and we love it

[AI/ML API @aimlapi]

@Coder9420948594 Nerfed Fable is still good, but not really worth its price

[AI/ML API @aimlapi]

@n1ckstr3 masks were the funniest part

[AI/ML API @aimlapi]

@adxtyahq Yup, not too cost-efficient

[AI/ML API @aimlapi]

@notjazii K3 is absolutely great too, no questions on that

[AI/ML API @aimlapi]

@mihuq you're welcome

[AI/ML API @aimlapi]

Sure, here's one for the ship:

THE SEA (same for everyone):
- Ocean plane ~500 x 500 m, ~160 x 160 segments, displaced in the VERTEX SHADER by a sum of several directional sine waves. The sea is lively but not extreme: the biggest waves are about 2.5-3 m high with ~40 m wavelength, plus smaller chop. Slightly sharpened crests so waves look real, not sinusoidal.
- Deep blue water (#0e3550 range) that stays BLUE in daylight, fresnel sky reflection, sun glitter, white foam on steep crests, a foam wake ~25 m behind the stern, churned foam at the bow, light haze at the horizon.

ANIMATION PLAN (all motion, no user input):
- ONE wave-height function h(x, z, t) shared by the water shader and the JS ship physics. Sample at bow/stern/port/starboard each frame; ship height = average - draft 0.8 m; pitch and roll from the differences, smoothed ~0.3 s. The ship rides the waves - never detaches, waves never clip through the deck.
- ROWING: all 22 oars in one synchronized 2.2 s cycle with four phases - CATCH (blades swing forward and dip into the water with an 8-14 particle splash burst), DRIVE (~0.9 s underwater sweep aft), RELEASE (lift clear with droplets), RECOVERY (feather forward above the waves). Blade tips must reach the actual water surface height. Smooth easing, no snapping.
- SAIL: strong belly toward the bow; a gust every ~7 s deepens the belly ~15% and releases; constant light ~9 Hz flutter along free edges and brails; the yard tilts a few degrees with gusts and roll; rigging follows its anchors.
- SPRAY: pooled particles (<= 1000, pre-allocated): bow bursts when the bow plunges, oar splashes, gravity, sub-second fade.
- SECONDARY LIFE: steering oar works small corrections (~5 s rhythm) with its tiller; shields sway 1-2 degrees with phase offsets; barrels and amphorae tremble subtly. Nothing is perfectly still.

CAMERA (identical for everyone - mandatory, no controls):
- Fully automatic from the first frame: the camera circles the ship at radius 55 m, height 12 m, always looking at (0, 3, 0) - a balanced medium shot: the whole ship with mast and sail comfortably in frame, never cropped, never tiny. One full revolution in exactly 36 seconds, constant speed, counter-clockwise. FOV 45. Nothing else moves the camera.

RELIABILITY (mandatory - the result MUST run):
- No placeholders, no TODOs, no empty arrays passed to curves or geometries, no NaN, no undefined variables. Prefer the simple construction that definitely runs over the complex one that might break.
- Budget: under 250k triangles (instance repeated details), no per-frame allocations, no heavy CPU loops; smooth 60 fps on a normal laptop.
- SELF-CHECK before you finish - verify each of these in your own code:
1) If the ocean uses PlaneGeometry, it is laid FLAT (rotateX(-Math.PI/2)) and is actually visible from the camera.
2) There are NO '#extension' directives inside any shader string, and no reserved GLSL words (like 'patch') are used as variable names.
3) Every object you reference in the animation loop is defined and stored under EXACTLY the same name/key it was created with; no curve or geometry receives an empty or partially built points array.
4) The ship reads as a SOLID closed wooden vessel from every angle; the shields sit exactly on the gunwales; the oars never intersect each other or the hull.
5) The camera orbit runs by itself and the whole ship stays in frame for the entire revolution.
6) The page contains zero visible UI - canvas only.
7) All animations run: wave riding, 22 oars with splashes, sail belly + gusts + flutter, steering oar, shield sway.

GOAL: a bright, crisp, movie-grade daytime shot - the richly detailed Greek war galley of Odysseus shouldering through big blue seas, foam and spray flying, oars beating in rhythm, sail full - like a production still from an epic film. Detail and smoothness above everything.

[AI/ML API @aimlapi]

@hypersonq We think about it freQWENtly

[AI/ML API @aimlapi]

@n1ckstr3 march on the Troy

[AI/ML API @aimlapi]

@hypersonq We prefer Greece 🏦

[AI/ML API @aimlapi]

@n1ckstr3 @Alibaba_Qwen We'll take a look into that 👀