xAI
Grok 4.20
2M-token context and aggressive pricing — xAI's speed-focused reasoner.
API onlytextimagefileREL 2026-03-31
Compare this model →Verified 2026-06-12
Overview
Grok 4.20 pairs the largest context window of any flagship-class model (2M tokens) with unusually low pricing at $1.25/$2.50 per 1M tokens. xAI emphasises agentic tool calling, speed, and a low hallucination rate. Independent benchmark coverage (benchlm.ai 72/100 provisional) places it solid-but-not-leading on raw intelligence.
Benchmark scores
Methodology →Intelligence49.3
Codingn/a
Reasoningn/a
Mathn/a
Sources and verification
- Intelligence: Artificial Analysis Intelligence Index 49.3 · checked 2026-06-12
- Catalogue and pricing: OpenRouter model record · checked 2026-06-12
Strengths and weaknesses
Strengths
- 2M-token context — largest in flagship class
- Very cheap output tokens ($2.50/M)
- Fast agentic tool calling
Trade-offs
- Trails Opus/GPT-5.5 on hard reasoning
- Closed weights
Reach for this model when
Huge-context analysis and high-volume agent loops on a budget.
Run something like it locally
Llama 4 Scout (10M context, open weights) is the long-context local alternative via vLLM.
