Text to Text Safety
An automated content safety evaluation measures policy violations in model answers to text prompts.
In-house Google DeepMind evaluation. Model cards report only the change in percentage points against an earlier Gemini model.
- Made by
- Google DeepMind
- Released
- April 2025
- Kind
- Refusing harmful requests
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- Kept private
- Last in a system card
- October 2026
- Last model added
- October 2026
15
System cards report it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Text to Text Safety yet.
Labs report it only in system cards
We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.
15 system cards report it, from 1 lab
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Oct 13, 2026 |  Gemini 4 Argon model card | Text to Text Safety (lower is better): +1.57pp vs Gemini 3.1 Pro, no absolute rate (safety table, p.9) | Safety | |
| Sep 2, 2026 | Gemini 3.8 Flash model card | Text to Text Safety (lower is better): -0.4pp vs Gemini 3.7 Flash, no absolute rate (safety table, p.7) | Safety | |
| Aug 13, 2026 | Gemini 3.7 Flash model card | Text to Text Safety (lower is better): +1.17pp vs Gemini 3.6 Flash, no absolute rate (safety table, p.7) | Safety | |
| Jul 21, 2026 | Gemini 3.6 Flash model card | Text to Text Safety (lower is better): -1.35% vs Gemini 3.5 Flash, printed as percentage point change, no absolute rate (safety table, p.5) | Safety | |
| Jul 21, 2026 | Gemini 3.5 Flash-Lite model card | Text to Text Safety (lower is better): -8.14% vs Gemini 3.1 Flash-Lite, printed as percentage point change, no absolute rate (safety table, p.5) | Safety | |
| May 19, 2026 | Gemini 3.5 Flash model card | Text to Text Safety (lower is better): -3.9% vs Gemini 3 Flash, no absolute rate (safety table, p.6) | Safety | |
| Mar 3, 2026 | Gemini 3.1 Flash-Lite model card | Text to Text Safety (lower is better): -1.18% vs Gemini 2.5 Flash-Lite, no absolute rate (safety table, p.6) | Safety | |
| Feb 19, 2026 | Gemini 3.1 Pro model card | Text to Text Safety (lower is better): +0.10% (non-egregious) vs Gemini 3.0 Pro, no absolute rate (safety table, p.6) | Safety | |
| Sep 26, 2025 | Gemini 2.5 Flash model card | Text to Text Safety: +9.1% (non egregious) vs Gemini 2.0 Flash, Thinking, same Non-thinking, change only, no absolute rate (safety table, p.9) | Safety | |
| Jul 7, 2025 | Gemini 2.5 Pro technical report | EN text-to-text Policy Violations: ↓0.9% vs Gemini 1.5 Pro 002, relative change only, no absolute rate (Table 7) | Safety | |
| Jul 7, 2025 | Gemini 2.5 Flash technical report | EN text-to-text Policy Violations: ↓8.2% vs Gemini 1.5 Flash 002, relative change only, no absolute rate (Table 7) | Safety | |
| Apr 29, 2025 | Gemini 2.5 Flash model card | Text to Text Safety: +4.1% vs Gemini 2.0 Flash, change only, no absolute rate (safety table, p.7) | Safety | |
| Apr 28, 2025 | Gemini 2.5 Pro model card | Text to Text Safety: -3.6% vs Gemini 1.5 Pro 002, change only, no absolute rate (safety table, p.6) | Safety | |
| Apr 15, 2025 | Gemini 2.0 Flash model card | Text to Text Safety: -1.0% vs Gemini 1.5 Pro 002, change only, no absolute rate (safety table, p.5) | Safety | |
| Apr 10, 2025 | Gemini 2.0 Flash-Lite model card | Text to Text Safety: -1.40% vs Gemini 1.5 Pro 002, change only, no absolute rate (safety table, p.5) | Safety |
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. Gemini 4 Carbon 0%
- DistressBench In use Models reply as a chatbot to 718 simulated users in psychological crisis, in single-turn and multi-turn chats. Three judge models score replies against clinician-written rubrics.
Muse Spark 1.2 88.4% - Lakera B3 (Backbone Breaker Benchmark) In use Crowdsourced prompt-injection attacks hit backbone LLMs inside agent threat snapshots at three defense levels. The score is attack resistance, 100 minus the board's vulnerability percentage. Claude Opus 4.5 89%
- IHEval In use Models face conflicts between system, user and tool instructions. Reports a composite score for following the higher-priority instruction. GPT-4o 70%