docketrouter
Models / xAI

Grok 4.5

by xAI · x-ai/grok-4.5

Grok 4.5 is a model from SpaceXAI with frontier performance on coding, knowledge work, and STEM.

reasoningtool-usevisionreleased 2026-07-08
Legal score · raw
-
not yet benchmarked
Context
500K
max output 450K
Input
$2
per 1M tokens
Output
$6
per 1M tokens
Suite cost
-
run the suite to see

Benchmark results

TaskCategoryRawJuicedCorrectLatencyCostRan
Hearsay IdentificationEvidence------
Bluebook Citation FormatResearch & Writing------
Federal Civil ProcedureProcedure------
Limitations ArithmeticProcedure------
Contract Clause ClassificationContracts------
Citation Hallucination ResistanceReliability------