xAI Publishes Grok 4.5 Model Card: Benchmarks and Safety Evaluations
xAI has published the official model card for Grok 4.5, a technical spec sheet with the model's capability benchmark results and safety evaluations. Cards like this show what a model can do and where its limits lie before anyone starts using it. Grok 4.5 itself was unveiled on July 8, 2026. The release now comes with transparent technical documentation.
AI-processed from Cursor Blog; edited by Hamidun News
In July 2026, xAI published the official model card for Grok 4.5, a document that collects the model's capability benchmark results and the outcomes of its safety evaluations. The team unveiled Grok 4.5 itself on July 8, 2026, and the card adds detailed technical documentation to the release.
What is a model card
A model card is a neural network's spec sheet: a short, official document that a model's developer publishes to show what the model can do and where its boundaries lie. For xAI's Grok 4.5, the card brings together two sets of data: capability benchmark measurements and safety evaluation results.
- Model: Grok 4.5, the flagship version of the Grok family
- Grok 4.5 was unveiled on July 8, 2026
- Document: the official model card from the development team
- Contents: capability benchmarks and safety evaluations
- Format: a public document released alongside the model
What the card documents
The Grok 4.5 card records two types of results: measurements on standardized capability benchmarks and the outcomes of safety checks. A benchmark is a set of tasks with known correct answers, used to measure how well a model reasons, writes code, and works with text; these numbers let readers compare a new version with its predecessors and with competitors. The second block, the safety evaluations, shows how the model behaves in risky scenarios and which restrictions its developers built in.
"We are publishing the official model card for Grok 4.5.
It documents the model's capability benchmarks and safety evaluations," the team says in the blog post accompanying the release.
Why publish safety evaluations
Publishing safety evaluations alongside benchmarks shows not only a model's strengths but also its weak spots, before anyone starts using it. Without this section, users see only the strong points and don't know where the model may make mistakes or behave unpredictably. Releasing model cards has become an industry standard: major AI labs do it so that developers and regulators can assess risks in advance. For Grok 4.5, both sets of data, capabilities and safety, are collected in a single document published in July 2026.
What this means
The model card moves the Grok 4.5 announcement from the marketing realm to the verifiable one: instead of generic talk about "the most powerful model," there are concrete capability measurements and risk assessments. For teams building products on Grok, it is the reference document: they use it to decide whether the model fits a specific task and which of its answers can be trusted.
FAQ
What does the Grok 4.5 model card show?
The Grok 4.5 model card shows two sets of data: the model's results on capability benchmarks and the outcomes of its safety evaluations. It is an official document published alongside the release in July 2026.
When was Grok 4.5 unveiled?
Grok 4.5 was unveiled on July 8, 2026. The official model card with benchmarks and safety evaluations was released as part of that launch.
- Oct 28, 2026Wired Tested 6 AI Agents in a Day: Dots Leads, Gemini Spark Fails at Lunch
- Oct 28, 2026Meta releases Muse Spark Open: 816B parameters and the lead among US open-weight models
- Oct 28, 2026Goodfire: Agentic Analysis Decodes Neurons Once Written Off as Chaos
- Oct 28, 2026Cognition releases FrontierCode 2.0: the best score is just 45.5%
Key Questions Answered
What does the Grok 4.5 model card show?
The Grok 4.5 model card shows two sets of data: the model's results on capability benchmarks and the outcomes of its safety evaluations. It is an official document published alongside the release in July 2026.
When was Grok 4.5 unveiled?
Grok 4.5 was unveiled on July 8, 2026. The official model card with benchmarks and safety evaluations was released as part of that launch.
What is a model card?
A model card is a neural network's spec sheet: a short, official document that a model's developer publishes to show what the model can do and where its boundaries lie. For xAI's Grok 4.5, the card brings together two sets of data: capability benchmark measurements and safety evaluation results. Model: Grok 4.5, the flagship version of the Grok family Grok 4.5 was unveiled on July 8, 2026 Document: the official model card from the development team Contents: capability benchmarks and safety evaluations Format: a public document released alongside the model
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.