DeepSeek V4.1 Flash tested: 300-400+ tok/s, cheap and fast, but prone to overthinking
WorldofAI · youtube · 2026-09-09
- YouTuber WorldofAI ran a full hands-on test of DeepSeek V4.1 Flash across coding, 3D simulations, Three.js environments, Minecraft and Mario Kart-style games, rocket simulations, autonomous dungeon games, exploded camera views and vision tasks.
- The model hits roughly 300-400+ tokens per second in some tests, delivers surprisingly strong results at a very competitive price, and feels like a major improvement over previous DeepSeek Flash releases.
- Caveats: it can overthink, spends too much time testing, and occasionally struggles with instruction following.
Related event: DeepSeek V4.1 Flash Tested: Blazing 350 Tokens/s but Still Experimental(2 posts)→
More from Models
- GPT-6 'Astra' Does 34 Math Steps in Latent Space, 4x More Than Sol — MaartenBaert · 2026-09-10
- NVIDIA details Alpamayo 2 Super, its L4 autonomy model for robotaxis — drmapavone · 2026-09-10
- OpenAI users report usage quotas wiped to zero as weekly reset dates shift by two days — ___Patrice___ · 2026-09-10
- Prediction: V4.1-Flash to score 36-38 on new AA index, agency at 42 — teortaxesTex · 2026-09-10
- Perplexity benchmarks 13 retrieval models: pplx-embed-v1-4b leads two of three categories — perplexity_ai · 2026-09-10
- Meme Budget: 'GPT-8 Swarm' Eats 182M GPUs as Astra Weighs Pausing Pro Signups — burny_tech · 2026-09-10