Grok 4.6
SpaceXAI · 4 thinking configurations
Intelligence35.1
Input / 1M$2.00
Output / 1M$6.00
Context500K
Tariff: SpaceXAI · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost4 measured configurations · 0 without both values
Selected: thinking low · Click a point to switch configuration
Grok 4.6 is a model from SpaceXAI with frontier performance on coding, knowledge work, and STEM. It is succeeded by Grok 4.7.
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
35.1 pointsDeepSWE · pass@1Software-engineering tasks on mini-swe-agent v1.1. Pass@1 across evaluated runs, with measured cost per task. The 95% intervals may overlap.
41.6 % ±2.3GPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
87.9 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
27.6 %SciCodeCode generation for scientific research problems.
49.4 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
75.3 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
43.3 %Specifications from the catalog
Inputtext, image, file
Outputtext
Tool callingYes
Structured outputYes
Speed64.6 tok/s
First token8.22 s
Maximum output450K
Catalog listing date2026-08-12
DeepSWE v1.1 · 113 tasks · mini-swe-agent · cost $1.04 / task · 44 mean steps · 16.5K mean output tokens. Published by the source 9/22/2026.