Google DeepMind pilots the first ‘double-blind’ evaluation of a proprietary frontier model with Singapore’s AISI and MLCommons
Confirmed
The takeaway
On Aug 27, 2026 Google DeepMind described what it calls the world’s first double-blind evaluation of a proprietary frontier-class model.
Status
- Claim
Confirmed
- Our reporting
- High confidence
- Importance
- 3 of 5
- Last verified
- 30 September 2026
Your AI and this story
- GPT-6.2 Astra88 days after its cutoff
- Claude Opus 5.558 days after its cutoff
- Gemini 4 Hydrogen149 days after its cutoff
- Grok 4.788 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 58 days before it.
Key facts
- Partners: Singapore AI Safety Institute, OpenMined, AVERI and MLCommons (DeepMind)
- Uses Google Cloud Confidential Space to cryptographically verify that the evaluation data and the proprietary model stay private to their owners; The New Stack reports the run used an NVIDIA H100 confidential GPU instance
- Model tested: ‘a Gemini Flash Lite model’ (DeepMind); The New Stack identifies it as Gemini 2.5 Flash Lite, run against reserve prompts from MLCommons’ AILuminate safety benchmark and a separate private prompt set for the Singapore context
- Removes the old trade-off in which evaluators either hand over test prompts (contamination risk) or labs hand over weights (IP risk); DeepMind says it matters most for sensitive cyber and government evaluations
What happened
External safety testing usually forces a trade-off. Either the evaluator gives the lab its test prompts, which can then leak into training and inflate scores, or the lab gives the evaluator its model weights, which it does not want to do. Google DeepMind and four partners ran an evaluation where neither happened. Inside a hardware-isolated, attested Google Cloud environment, the evaluators’ confidential benchmarks ran against a Gemini Flash-Lite model: “The evaluator cannot see the Gemini model weights, and Google cannot see the evaluator’s test prompts.” DeepMind published its methodology and findings with the announcement.
Why it matters
As AI safety institutes and outside auditors gain a formal role (e.g. California SB 813’s independent assessors, OpenAI’s Sept 22 principles for third-party assessments), a trusted way to test closed models on secret benchmarks becomes basic infrastructure. The pilot used a small model; it is not yet shown to scale to frontier models.
Sources
3 sources from 3 sites. Numbers match the chips in the text.
3 sources: 1 primary, 2 press
Primary
- Google DeepMind: Piloting the world’s first double-blind AI evaluationsdeepmind.google, official
Press
Changes
- Filed