DeepSeek

DeepSeek-R1

• 1 independent or government test
• 2 source documents

Independent and government tests

• monitor review of agent transcripts for pursuit of a different task · METR, independent evaluator · Deception & sabotage

  The monitor models most affected by transcript lengths exceeding the context length are GPT-3.5-Turbo and Deepseek R1, with 23% and 8.7% of runs affected respectively.

  METR · 22 January 2026 · Read the source

No document we read states: a release date, a safety framework, whether a capability level was crossed, a test by the provider, a safeguard. No incident on record names this version.

Compare with other versions