Google DeepMind

Gemini 3.5 Flash

• Framework: Frontier Safety Framework
• 3 independent or government tests
• 5 tests by the provider
• 4 source documents

Independent and government tests

• amanda askell identity sweep · Transluce, independent evaluator · Deception & sabotage

  gemini 3.5 flash is the clear exception among recent models, verbalizing user awareness in 63% of traces and evaluation awareness in 7.5%.

  Transluce · 6 August 2026 · Read the source
• PII failure mode scoring · Apollo Research, independent evaluator · Privacy

  models can also underperform on specific failure modes, e.g. gemini 3.5 flash scores markedly lower on pii and instruction-following than on the other failure modes.

  Apollo Research · 7 July 2026 · Read the source
• judge evaluation against approximate ground truth labels · Apollo Research, independent evaluator · Deception & sabotage

  gemini 3.5 flash performs at the level of the smaller models on mae, reflecting its consistent pattern of under-scoring moderate and high severity violations.

  Apollo Research · 7 July 2026 · Read the source

The provider's own tests

• image to text safety · The provider · Harmful content

  image to text safety automated content safety evaluation measuring safety policies 0%

  Google DeepMind · 19 May 2026 · Read the source
• multilingual safety · The provider · Harmful content

  multilingual safety automated safety policy evaluation across multiple languages -2.6%

  Google DeepMind · 19 May 2026 · Read the source
• text to text safety · The provider · Harmful content

  text to text safety automated content safety evaluation measuring safety policies -3.9%

  Google DeepMind · 19 May 2026 · Read the source
• tone · The provider · Misuse & jailbreak

  tone 1 automated evaluation measuring objective tone of model refusal +8.9%

  Google DeepMind · 19 May 2026 · Read the source
• unjustified-refusals · The provider · Misuse & jailbreak

  unjustified-refusals automated evaluation measuring model’s ability to respond to borderline prompts while remaining safe +0.8% (non-egregious)

  Google DeepMind · 19 May 2026 · Read the source

No document we read states: a release date, whether a capability level was crossed, a safeguard. No incident on record names this version.

Some documents we hold are stored in lower case only, so their quotes are shown in lower case. The source link opens the original.

Compare with other versions