Builds deterministic ~100k/200k token prompts, calls an OpenAI-compatible /v1/chat/completions endpoint and records status, latency, usage and errors. Intended for evaluating free-tier limits under long-context novel-continuation workloads.

Required Ruby Version

>= 3.0

Authors

Alex Warren

Versions

  1. 0.1.1

    1. source

      1. 0.1.1 source October 22, 2026 (8 KB)
  2. 0.1.0

    1. source

      1. 0.1.0 source October 22, 2026 (8 KB)

Pushed by

longctxbench

SHA 256 checksum

← Previous version