Grok 4.7
xAI (SpaceXSI) · Multimodal model
Grok 4.7 is xAI's current flagship model, released September 21, 2026 at the same price as Grok 4.6. It accepts text and images, has a 500K-token context, and targets coding, agents and knowledge work. A faster variant is offered at…
Current
ReleasedSep 21, 2026
Context window500,000 tokens
Input → outputtext, image → text
API price$2.00 / 1M input, $0.50 / 1M cached input, $6.00 / 1M output (prompts < 200k tokens); $4.00 input, $1.00 cached, $12.00 output (prompts ≥ 200k tokens, applied to all tokens in the request).
Consumer access—
Best for
Coding, agentic tasks and professional knowledge work (e.g. legal) at a mid-range price.
Overview
Grok 4.7 is xAI's current flagship model, released September 21, 2026 at the same price as Grok 4.6. It accepts text and images, has a 500K-token context, and targets coding, agents and knowledge work. A faster variant is offered at twice the price.
Strengths and limits
- Big jump over Grok 4.6 on Terminal-Bench 4.0 (37.6% vs 20.3%)
- Tops xAI's comparison on Harvey Legal Agent Benchmark (19.6%) and EEBench
- Same price as Grok 4.6; reasoning effort up to 'xhigh'
- Trails Claude Fable 5.1 on coding rows in xAI's own table
- No Batch API; context window smaller than Grok 4.3's 1M
Published benchmark scores
Terminal-Bench 4.0 (xHigh)37.6%
Source: x.ai
DeepSWE v1.1 (high effort)71.0%
Source: x.ai
CursorBench 4.0 (xHigh)46.3%
Source: x.ai
Harvey Legal Agent Benchmark (xHigh)19.6%
Source: x.ai
Scores are as published by the source; settings (reasoning effort, tools, attempts) differ between labs, so compare with care. How to read benchmarks →