# Grok 4.1 Fast on the Cotera agent benchmark

> Calls the tools, reads the responses, and then forgets to write anything down. 1/5.

Source: https://cotera.co/benchmarks/models/grok-4.1-fast

---

xAI's Grok 4.1 Fast — the cost-tier variant in the Grok 4.1 family. Tested via xAI's first-party API.

**Score:** 1/5 benchmarks passed · **Cost across the matrix:** $0.089

## Strengths
- Sales benchmark passed cleanly: 6 tool calls, $0.035. Lowest-cost Sales pass in the matrix that involved actual verification.

## Watch-outs
- Empty final message after tool calls on Marketing (7 tools), CX (9 tools), and Apollo (0 tools). Pattern is: Grok runs the agent loop, processes tool responses, then closes the run without composing the final answer.
- Coding was 0 tool calls and no output — Grok didn't even attempt to engage with the task.
- Total spend was $0.090, but four out of five runs produced nothing. You're paying for tool retrieval the model couldn't be bothered to write up.