Design Arena

AGI-01 Swift by @LucidQuery is now available on Design Arena! Built for everyday work, AGI-01 Swift works quickly to complete writing, coding, search, and reasoning tasks effectively and efficiently. AGI-01 Swift runs on LucidQuery’s World Engine system to multiply the effective capability of the model. Congrats to the @LucidQuery team on the launch!

UNI-1.1 Max by @LumaLabsAI is 8th on Image Arena on Design Arena with an Elo of 1249. This puts it in the same performance band as Gemini 3 Pro Image Gen 2K (Nano Banana Pro) by @GoogleDeepMind. UNI-1.1 Max ranks 6th in Graphic Design Arena with an Elo of 1289, and 6th in Logo Arena with an Elo of 1272, following closely behind Gemini 3.1 Flash Image Gen 2K (Nano Banana 2) by @GoogleDeepMind in both categories. Congratulations to the @LumaLabsAI team on these achievements!

Interestingly, all of GLM-5.2’s games look the same. GLM-5.2 games have a neon/cyberpunk aesthetic with CSS gradients, neon glow effects (text-shadow, box-shadow), and CSS animations/transitions. Here is a UMAP projection of 1000 randomly sampled GLM-5.2 game dev generations, grouped by similarity:

Prompt (abbreviated): Create a polished, high-quality 3D arcade racing game called "Race Arena”

Some key patterns we saw in generated games: 1. Audio is baked in almost universally (Tone.js synths were used for SFX) 2. Canvas 2D is the dominant renderer (over both DOM manipulation and WebGL) 3. Self-rolling physics (custom gravity/velocity/bounce rather than using physics libraries) 4. Heavy visual polish (particles, screen shakes, gradients, and glow effect are dominant) Here a few gameplay examples showcasing GLM-5.2’s game dev capabilities:

GLM-5.2 by @Zai_org is 2nd on Game Dev Arena on Design Arena with an Elo of 1368. This is a 6 position and 29 Elo jump from GLM-5.1, putting GLM-5.2 in the same performance band as Claude Fable 5 by @Anthropic. GLM-5.2 is the top open weight lab in Game Dev and second lab overall, ahead of @OpenAI and just behind @Anthropic. Congratulations to the @Zai_org team on this achievement!

Explore more leaderboards at DesignArena.ai

With an input cost of $0.015, Krea 2 Turbo establishes a new Pareto frontier in Image Preference vs. Price, alongside Krea 2 Medium, GPT Image models by @OpenAI, and Z-Image Turbo by @Alibaba_Qwen.

Krea 2 Turbo by @krea_ai is 13th on Image Arena with an Elo of 1234. This is in the same performance band as MAI-Image-2.5 by @MicrosoftAI. Optimized for speed, Krea 2 Turbo is 16.5 seconds faster than Krea 2 Large and improves the most in typography and product Image categories. Congratulations to the @krea_ai team on these achievements!

HappyHorse 1.1 by @AlibabaGroup is now available on Design Arena! HappyHorse 1.1 introduces improvements in visual and motion expressiveness, character and text stability, and prompt adherence to enhance quality, control, and efficiency in professional content creation scenarios. Congrats to the @HappyHorseATH team on the launch!

Step 3.7 Flash establishes @StepFun_ai as the top 8 open weight lab on Design Arena, behind @Alibaba_Qwen and @NexEcosystem. @StepFun_ai’s latest model, Step 3.7 Flash, ranks 22nd among open weights with an Elo of 1216. It excels at creating social media websites, blogs, and corporate websites. Congratulations to the @StepFun_ai team on this achievement!

@mathemagic1an We’re delighted you like them! We’re all ears for more suggestions :)

@Designarena These posts are an act of public service, thank you

@xeophon Thank you! We might need your help here @FireworksAI_HQ

@jeremyphoward @Zai_org We agree! We even found it to be better than Opus 4.8, GPT 5.5, and all other models at frontend: x.com/designarena/st…

Wow. @Zai_org GLM 5.2 is a marvel! It is *at least* as good as Opus 4.8 and GPT 5.5. It's super fast, inexpensive, and not too verbose. It responds with nuance and judgement, & handles long context VERY well. I've never experienced an open weights model like this before.

@ZixuanLi_ We're looking forward to seeing how it performs in Mobile Arena (React Native and Android)!

GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks. Results: - GLM-5.1: 21/70 - GLM-5.2: 48/70 - Claude Fable 5: 56/70 That's more than a twofold improvement from GLM-5.1 to GLM-5.2. These come from an internal benchmark of 35 challenging mobile development tasks, each run twice for a total of 70 trials. We measured task completion, defined as core features working without major issues.

Kimi K2.7 Code by @Kimi_Moonshot is 5th overall among open weight models on Design Arena with an Elo of 1312. This is in the same performance band as MiniMax M3 by @MiniMax_AI. With an average generation time of 337.6 seconds, Kimi K2.7 Code is 78.8 seconds faster than Kimi K2.6 on average. @Kimi_Moonshot is among the top 2 open-weights AI model labs and top 3 labs overall on Design Arena. Congrats to the @Kimi_Moonshot team for this accomplishment!