Which one feels closest to the real thing?

If you had to choose one model to build with, which one would you pick? One game concept. Four frontier models. One Temple Run-inspired 3D endless runner. GPT-6 Astra — 4,033 tokens, $1.280423 GLM-5.3 — 4,337 tokens, $0.018591 DeepSeek

Novita AISep 115

Devlog #3 Nanoharness I'm not sure if I calculate TPS correctly

But deepseek V4.1 flash seems incredibly fast despite only 100t/s Also I think I need to work on cache hit, or check the calculation again

balega_devSep 11

Atomic Chat ran a test: feed DeepSeek-V4.1-Flash a prompt

Output a single local HTML file, and ask for a ray-traced FPS with weapon mechanics and enemy waves. It worked. No asset pipeline. No UI framework. One pass. The model is a 552B MoE with a Causal-Encoder-Decoder and

Pascual ⚡Sep 11

This Anthropic report release was pretty wild

Especially the section where they mentioned how Kimi and DeepSeek routes API calls to Opus model sometimes. I have some thoughts, and showing you encrypted reasoning traces that Anthropic says Chinese labs have been stealing.

Mehul MohanSep 1123