Grok 4.1 Fast

Grok 4.1 Fast is not yet on DualEntry's published leaderboard, so our accounting-task scores are **synthesized** from adjacent benchmarks, public sentiment, and editorial judgment — we estimate ~58% overall, which would place it in the lower-middle of the roster on capability.

What Grok 4.1 Fast does dominate: cost ($0.20 / $0.50 per MTok), context (2M tokens — double any other model on the index), and price/capability ratio. Prior editions listed a ~200 tok/s output speed; Artificial Analysis's measured number is closer to 135 tok/s, corrected here. Even at the measured speed, Grok 4.1 Fast is among the fastest models indexed at this price tier.

For agentic workflows where the accounting task is narrow enough that ~60% accuracy on structured sub-tasks is acceptable with a human reviewer — and where the context window matters for ingesting years of ledger history — Grok 4.1 Fast is a legitimate option that deserves more serious evaluation than its brand position usually gets in finance circles. xAI's more capable Grok 4.20 (indexed separately) is the capability-upgrade path for buyers already familiar with this family; Grok 4.1 Fast remains the correct choice for absolute cost-efficiency. Update when DualEntry publishes a measured number.