DeepSeek Bench Scores Rely on Max Reasoning, Inflating Task Time by Up to 10x
TheZachMueller · x · 2026-08-01
Community members note that DeepSeek's recently announced high benchmark scores were all achieved using max reasoning effort.
In this mode, the model consumes 2 to 10x more output tokens at a slightly lower drafter acceptance rate. Consequently, the actual wall time to complete tasks inflates to 5 to 10x compared to normal conditions. For instance, a task that typically takes 5 minutes could take 25 to 50 minutes under max reasoning, explaining why users might experience significant latency in practice.
More from Models
- Users report Gemini's memories can't actually be deleted even after wiping chat history — ThrowRA_Purpose2025 · 2026-10-16
- OpenAI stealth-testing mysterious internal model 'Aeon' in Codex desktop app and Dots — cedric_chee · 2026-10-16
- Argon edges out astra as daily driver, but Claude still wins heavy coding work — iruletheworldmo · 2026-10-16
- NVIDIA ships FP4-quantized Qwen3.8-Flash for local use, trending on Hugging Face — nvidia · 2026-10-16
- RoboVista benchmark: Gemini-4 Argon hits 81.9% on robotics understanding — berkeley_ai · 2026-10-16
- Dev reports decision model jev beats gpt-6-luna on quality in real training pipeline — genmon · 2026-10-16