First Community Benchmarks of DeepSeek V4 Pro, Qwen 3.8 Max, and Grok 4.6

The community quickly tested three newly released LLMs—DeepSeek V4 Pro, Qwen 3.8 Max, and Grok 4.6—across 3D scene generation, Flappy game creation, and airplane‑animation tasks, comparing quality, speed, and cost to reveal each model’s strengths and trade‑offs.

PaperAgent
PaperAgent
PaperAgent
First Community Benchmarks of DeepSeek V4 Pro, Qwen 3.8 Max, and Grok 4.6

LLM development is exploding, with three cutting‑edge models launched in a single day, offering benefits for both consumers and enterprises.

Grok 4.6 (by Musk): performance comparable to Fable 5 while costing about 85% less.

Qwen 3.8‑Max : 2.4 T parameters, 95 B activation parameters; weights released on Hugging Face ( https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B).

DeepSeek‑V4‑Pro‑0813 : strong performance that goes beyond simple benchmark scores.

3D Scene Generation

Comparison: Grok 4.6 vs DeepSeek V4 Pro vs DeepSeek V4 Flash vs GPT‑5.6 Sol.

The gap in raw quality is minimal; Grok 4.6 and DeepSeek V4 Pro are very close. When speed is considered, the author would choose Grok 4.6. However, both lag noticeably behind Kimi K3, Opus 4.8, and Qwen 3.8‑Max.

https://x.com/OmedVibeCodes/status/2087660331761811533

Flappy Game Benchmark

Using FlappyBench, the three models—DeepSeek V4 Pro 0813, Kimi K3, and GLM 5.2—were tested with identical prompts and the /design command.

Scoring considered gameplay functionality, UX/UI, and cost:

DeepSeek V4 Pro 0813 → 8/10 · $0.0005 per game

Kimi K3 → 9.5/10 · $0.0740 per game

GLM 5.2 → 9/10 · $0.0480 per game

DeepSeek’s price is only 1/148 of Kimi’s and 1/96 of GLM’s, yet it achieves about 84% of their quality, making $0.0005 for a playable game seem astonishingly cheap.

Kimi K3 delivers the best output quality, with GLM 5.2 close behind.

https://x.com/CommandCodeAI/status/2087631884406939727

Airplane Sliding Animation

Grok 4.6 and DeepSeek 4 Pro were freshly released and tested for generating a basic airplane‑on‑water sliding animation, with results compared against DeepSeek 4 Flash and Fable.

Only the raw models were used, without any auxiliary frameworks.

Grok’s plane moves backward.

DeepSeek 4 Pro moves backward after one turn, then forward in another.

DeepSeek 4 Flash follows the correct direction but adds its own style.

Fable remains the clear winner.

Grok’s visuals are impressive, yet it still struggles with realistic physics, and further improvements are hoped for.

https://x.com/Terry_Djony/status/2087729113578758504
Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

AILLMDeepSeekbenchmarkQwencost analysisGrok
PaperAgent
Written by

PaperAgent

Daily updates, analyzing cutting-edge AI research papers

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.