SK retweeted
Introducing Gemini 4 Carbon, it's a big jump over Argon which we released just 2 weeks ago.
Gemini 4 Carbon is SOTA on cyber, long horizon tasks, ranks #1 on Artificial Analysis V5, significantly improves on usability in domains like coding, and is more aligned.
Here is Deepseek thought on the AAI V5 rankings:
My Assessment: Not Proven Gaming, but a Benchmark You Should Discount
I do not believe there is public evidence of a direct instruction to "make Anthropic and OpenAI look best." The private-test shift has a defensible technical rationale, and the top rankings are broadly consistent with other evaluations.
But the question is not whether there is a smoking gun. The question is whether the leaderboard is structurally trustworthy in the specific moment when trust matters most. It is not. A private evaluator, funded partly by the AI industry, selling services to the ranked labs, releasing an unauditable methodology days before the largest AI IPOs in history, producing results that widen the closed-open gap dramatically—this is not a neutral measurement. It is a valuable but compromised signal.
Treat the v5 leaderboard as one data point among many, and weight it accordingly. For the open-weight ecosystem especially, its verdict should carry little independent authority until the private test sets are opened to third-party audit or the business model is separated from the ranking function.
@teortaxesTex @TheAhmadOsman @Tech2Wild
#ai #Leaderboard #LLM
I hope the next @deepseek_ai and @Alibaba_Qwen models starts focusing broadly on things like science, physics, maths, but just heavily focused on coding and fancy fronted design.
SK retweeted
PROMPT
Build a single-file HTML page with three.js (ES module from jsdelivr, [email protected]) that renders a photoreal, fully procedural human eye in the browser. No textures, no 3D models, no audio files - everything generated in shaders and code.
SCENE
- Eyeball: unit sphere with a custom ShaderMaterial.
- Sclera: warm off-white, pinker toward the edges, thin red veins from ridged fbm noise, pink caruncle in the inner corner.
- Cornea: refract the view ray (IOR 1.376) onto a recessed iris plane so the iris shows parallax and depth.
- Iris in polar coords: radial fibres (multi-octave noise), crypts (dark lens-shaped holes), a wavy collarette ring, a dark limbal ring, a pupillary ruff and pigment freckles. Use the iris height field for bump lighting, and add a caustic on the side opposite the light.
- Wet surface: a rectangular softbox reflection on the cornea, a tiny sharp glint and Fresnel environment reflection.
- Skin patch: an irregular blob-shaped sheet with a rounded thick rim, built from rings that run from an almond-shaped lid opening out to the edge, and rebuilt every frame on the CPU.
- The lids hug the eyeball (smooth-max against the sphere), with an upper lid crease and a wet pink lid margin.
- Skin shader: per-channel wrap lighting for subsurface scattering (red wraps furthest), pores from voronoi, fine crinkles parallel to the lids, crow's feet at the outer corner, freckles, mottling, an oily specular and a peach-fuzz rim.
- Lid shadows on the eyeball computed analytically from the same almond function, plus striped lash shadows and a tear meniscus highlight along both lid margins.
- About 500 eyelashes as camera-facing tapered ribbons: curled, clumped into groups, 2–3 rows, longer in the middle. Upper lashes rotate downward when the eye closes.
- Background: warm brown gradient with soft diagonal light streaks and a contact shadow under the patch.
BEHAVIOR
- The eye follows the mouse cursor. When the cursor is idle, it makes random saccades with micro-tremor.
- The upper lid follows vertical gaze. Natural random blinks every 2–6s (fast close, slower open).
- Pupil hippus (slow breathing of size); it constricts after a camera flash and slowly recovers.
- Clicking the eye grows a new random iris that crossfades from the old one. There are 8 base palettes (amber, hazel, green, blue, grey, teal, brown, ice) with jitter, and an occasional golden central heterochromia.
POST
- EffectComposer with 4x MSAA, BokehPass focused on the cornea, a subtle UnrealBloom on highlights, and a final pass with chromatic aberration, vignette, film grain and a light sharpen. Tone-map with ACES.
UI (HTML overlay, monospace, uppercase, warm cream text with a yellow accent)
- Top-left "eye exam" panel: iris swatches + "new" button, pupil slider with auto toggle, 6 skin tones, lash length, framing (macro / portrait / wide / space - in space the patch floats and rotates), focus (shallow / sharp), blink and squint buttons, volume.
- Corner crosshair marks, slowly rotating thin dashed HUD rings, side readouts for gaze in degrees and pupil in mm, top-right resolution / fps / iris name.
- Snapshot button: white flash, the pupil constricts, and a thumbnail of the frame appears.
- Drag to orbit, scroll to zoom.
SOUND (Web Audio API only, starts on the first click)
- Ambient: a soft low drone (D3, A3, D4) plus a generative pad of sine notes picked from the D major pentatonic every 2–5s, with long attacks, convolution reverb (generated impulse response) and a feedback delay.
- UI tick on every action: a 15ms high-passed noise click plus a short pitched sine blip from the scale. Quieter ticks on hover.
- Blink: a soft low pitch-drop thump. New iris: a 4-note rising arpeggio plus a filtered noise sweep with tremolo. Framing change: a tremolo noise swell. Snapshot: a two-stage shutter click.
- Footer sound on/off toggle with a small live level meter.
Keep it 60fps on a normal laptop, make it work at phone width, and put everything in one
I must say @Kimi_Moonshot code plan limits has improved allot from the initial launch.
Again @OpenRouter the @deepseek_ai V4.1-flash-1017 AAI score your have is incorrect.
I think @Kimi_Moonshot should keep the Kimi-k3.1 pro closed and not release the weights
I know @Kimi_Moonshot coding plan sucks. But Mimi-k3.1 pro with max thinking effort is a solid. I’ll say if you wanna try it out just add credit or use openrouter.
#kimi #moonshot #ai #llm
So I think @Kimi_Moonshot fixed the quota limit issue for k3.1-Pro. If it’s a bug they need to be proactive in letting the community know and send out a reset card, that’s something moonshot still don’t know how to do. They needs a proper PR people.
I’ve seen enough , I don’t think average AI users need anything more than @deepseek_ai V4.1-flash 1017 at this point. That price and intelligence is just too good. Kimi-k3.1 pro might seem cheaper than the big labs, but its token hungry and overall expensive to use
Now I truly believe @Alibaba_Qwen 4Max will surpass Kimi-K3.1 Pro. Qwen4max is a newer architecture and it’s way bigger than Kimi-K3. This could easily be 60 plus on AA Intelligence.
I like the new @Kimi_Moonshot K3.1-Pro, but had to cancel my pro plan. The limits are just not worth it, I’ll rather use via openrouter when I want.
#kimi
Now I’m really looking forward to @Google Gemini Carbon, @Zai_org GLM-5.5, and @Alibaba_Qwen Qwen4-Max
China now have fable level AI way sooner than people thought . Thanks to @Kimi_Moonshot
Seems @deepseek_ai is gonna be our only saving grace after all. No matter how good @Kimi_Moonshot is, their coding plan is just plain trash and a rip off.





