Grok-1.5

Grok-1.5

Quick answer: Grok-1.5 (released March 28, 2024) was xAI's second major model, significantly improving on Grok-1's reasoning and coding capabilities with a 128K-token context window. It was the first Grok model with real-time access to X (Twitter) data, scoring 50.6% on MATH and 90.0% on HumanEval at launch.

At a Glance

Where Grok-1.5 leads

• Real-time X (Twitter) knowledge — unique access to live social data unavailable in other models
• 128K context window for long document and code analysis
• Competitive coding at HumanEval 90.0% at launch
• Strong performance for its generation on math: 50.6% MATH benchmark

Where it lags

• Superseded by Grok-2 and Grok-4 for production use
• Coding benchmarks (SWE-bench) below the 2025–2026 frontier
• API access limited; primarily integrated into X Premium

Best for: X-native applications requiring real-time social context, legacy integrations, and teams that pinned to this version.

What Grok-1.5 Is

Grok-1.5 was released on March 28, 2024, marking xAI's first significant capability leap since the initial Grok-1. The model was built with xAI's custom training infrastructure and showed particularly strong improvements in reasoning and math relative to Grok-1. It introduced the 128K context window and was made available to X Premium subscribers before a wider API rollout.

xAI's unique positioning in the LLM market is real-time access to X (Twitter) data — Grok-1.5 could reference recent posts, trending discussions, and public figures' statements in ways that knowledge-cutoff models could not. This remains the defining differentiator of the Grok line.

Specifications

Pricing

Public Benchmark Scores

Scores reported by xAI. Not Benchgen measurements.

Related Models & Benchmarks

• Compare: Grok-2 · Grok-4
• Benchmarks: MMLU-Pro

Specs sourced from xAI — Grok-1.5. Last updated 2026-06-19.

Grok-1.5

Grok-1.5 is a large language model developed by xAI.