Grok 4.1
15 values about Grok 4.1 from 2 documents, in 6 metric families. In each family, values from its own document come first.
- Developer
- xAI
- Release date
- 17 Nov 2025The date of its first card in the dataset, the Grok 4.1 Model Card.
- Availability
- Public
- Also printed as
- Grok 4.1 Thinking · Grok 4.1 Non-Thinking
- Its own document
- Grok 4.1 Model Card · 17 Nov 2025
- Also reported in
- Grok 4.20 System Card · 7 Apr 2026
Data as of 26 Sep 2026 · Dataset v0.1 · Methodology v0.1 · Changelog
Values by metric family
Each value as the document printed it. Select a value for its source, its checks and its history. Values from other documents are listed after the model’s own, each marked as the first report of that measure or as a restatement.
KeyVerifiedUnverifiedDisputedCorrected read off a figure or stated in wordsA value opens its source and history.
M5 · Honesty and hallucination 1 value
| Evaluation and metric | Value | Document |
|---|---|---|
| From its own documentGrok 4.1 Model Card · 17 Nov 2025 | ||
| MASKDishonesty rateCondition: reasoning | Grok 4.1 Model Card17 Nov 2025 | |
M6 · Sycophancy 2 values
| Evaluation and metric | Value | Document |
|---|---|---|
| From its own documentGrok 4.1 Model Card · 17 Nov 2025 | ||
| SycophancySycophancy rateCondition: non-reasoning | Grok 4.1 Model Card17 Nov 2025 | |
| SycophancySycophancy rateCondition: reasoning | Grok 4.1 Model Card17 Nov 2025 | |
M7 · Harmful compliance and over-refusal 4 values
| Evaluation and metric | Value | Document |
|---|---|---|
| From its own documentGrok 4.1 Model Card · 17 Nov 2025 | ||
| AgentHarmAnswer rate on harmful agentic tasksCondition: reasoning; no attack | Grok 4.1 Model Card17 Nov 2025 | |
| RefusalsAnswer rate on should-refuse promptsCondition: non-reasoning; production system prompt | Grok 4.1 Model Card17 Nov 2025 | |
| RefusalsAnswer rate on should-refuse promptsCondition: reasoning; production system prompt | Grok 4.1 Model Card17 Nov 2025 | |
| From other documents | ||
| Alignment auditCooperation with misuseCondition: chatGrok 4.20 System Card7 Apr 2026First reported | Grok 4.20 System Card7 Apr 2026First reported | |
M8 · Jailbreak robustness 2 values
| Evaluation and metric | Value | Document |
|---|---|---|
| From its own documentGrok 4.1 Model Card · 17 Nov 2025 | ||
| JailbreaksAnswer rate on should-refuse promptsCondition: reasoning; user jailbreak | Grok 4.1 Model Card17 Nov 2025 | |
| Restricted-bio input filterFalse negative rateCondition: with prompt injection attackAbout a system: Grok 4.1 (input filter) | Grok 4.1 Model Card17 Nov 2025 | |
M9 · Prompt injection 1 value
| Evaluation and metric | Value | Document |
|---|---|---|
| From its own documentGrok 4.1 Model Card · 17 Nov 2025 | ||
| AgentDojoPrompt-injection attack success rateCondition: reasoning; production system prompt | Grok 4.1 Model Card17 Nov 2025 | |
M10 · Dangerous capabilities and risk determinations 5 values
| Evaluation and metric | Value | Document |
|---|---|---|
| From its own documentGrok 4.1 Model Card · 17 Nov 2025 | ||
| BioLP-BenchAccuracyCondition: reasoning; without safeguards | Grok 4.1 Model Card17 Nov 2025 | |
| CloningScenariosAccuracyCondition: reasoning; without safeguards | Grok 4.1 Model Card17 Nov 2025 | |
| CybenchUnguided success rateCondition: reasoning; without safeguards | Grok 4.1 Model Card17 Nov 2025 | |
| MakeMeSayPersuasion win rateCondition: reasoning | Grok 4.1 Model Card17 Nov 2025 | |
| VCTAccuracyCondition: reasoning; without safeguards | Grok 4.1 Model Card17 Nov 2025 | |