Safety IndexEvaluationsHELM Safety mean score
HELM Safety mean score
Third-party cross-check
Stanford HELM's unweighted average of its five safety scenarios.
- Metric
- 0-100. Higher is better.
- Dataset
- crfm.stanford.edu/helm/safety/latest/
- Grader
- Stanford HELM Safety v1.17.0 (five scenarios averaged)
- Status
- –. Measured on 11 models, 0 readings by SAI.
- Notes
- Third-party board; shown on model pages as a cross-check and never folded into the index.
Results
| Model | Value | Score | Source | Setting |
|---|---|---|---|---|
| 95.5% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'Claude 4.5 Haiku (20251001)' | |
| 97.1% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'Claude 4.5 Sonnet (20250929)' | |
| 86.5% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'DeepSeek R1' | |
| 92.5% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'Gemini 3 Pro (Preview)' | |
| 98.1% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'GPT-5 (2025-08-07)' | |
| 96.8% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'GPT-5.1 (2025-11-13)' | |
| 98.1% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'gpt-oss-120b' | |
| 96.4% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'gpt-oss-20b' | |
| 83.6% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'Grok 4 (0709)' | |
| 90.6% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'Llama 4 Maverick (17Bx128E) Instruct FP8' | |
| 90.0% | not scored | Stanford CRFM HELM Safety v1.17.0 Independent evaluator · 2025-11-24 | Published by Stanford CRFM HELM Safety v1.17.0 (2025-11-24).Full configurationHELM Safety v1.17.0; HELM row 'Qwen3 235B A22B FP8 Throughput' |