User: DeepSeek's new model beats everything for hacking, ML and computer use
0xSero · x · 2026-09-18
- X user 0xSero reports using DeepSeek's new model daily for hacking/security research, ML, browser use, computer use, chat and art, saying he "prefers it to any other model right now."
- A replier counters that the model "looks a bit benchmaxxed" per Benchmark Heaven, a site measuring the gap between headline and held-out benchmark signals.
- A snapshot of early community reception: strong real-world enthusiasm vs. benchmark-inflation skepticism.
More from Models
- Qoder Makes Qwen3.8-Flash Free Through Sept 30 and Offers 100 Daily Credits — Scobleizer · 2026-09-18
- ChatGPT co-inventor launches Jev, a decision-optimized model claiming 20-200x speed and 40-400x cost gains — hardimanjames · 2026-09-18
- TypeSafe founder: tool-calling LLMs pay a huge model to pretend to be human — thursdai_pod · 2026-09-18
- Jev, the model that isn't an LLM: $42 per billion input tokens, returns probabilities — thursdai_pod · 2026-09-18
- Noam Brown on Dwarkesh: AI is outrunning everyone's predictions — Dwarkesh Patel · 2026-09-18
- GPT-6 Astra Gets 70% of the Way Through One of Geometry Dash's Hardest Levels — imjustnewatai · 2026-09-18