How Ressl launched Benchmark Router on X
Runs public LLM benchmarks on any model or your own harness
What Ressl launched

Introducing Benchmark Router - one place to run public benchmarks on any model / your harness. At a 1/10th of the cost to run yourself*. We have new LLM benchmarks launching... every 6 hours. Our goal with this platform is to give AI builders a unified place to run benchmarks https://t.co/ltpdv1GpHh
How LaunchVideo.com rated the launch
LaunchVideo.com rating
About the launch
Benchmark Router lets AI builders run public LLM benchmarks on a chosen model, or on an agent harness, without standing up each suite's official runner. You pick a benchmark, a model, and a configuration, from the site, the HTTP API, or the brouter command line. The service runs that benchmark's official code through a metered gateway and returns the official scores, a result row for every task, and the benchmark's own logs. A standard run is priced at 10% of the model API cost, with one credit equal to one dollar. A run on your own harness is billed at the full API cost. The live catalog covers knowledge, coding, agentic work, legal, finance, health, science, and vision, including Humanity's Last Exam, HealthBench, Terminal-Bench 4.0, Harvey's public Legal Agent Benchmark, SWE-bench variants, BankerToolBench, and EnterpriseOps-Gym. Each suite keeps the harness its leaderboard uses, from Harbor on Modal and mini-swe-agent to Claude Code, Codex, or the official CAIS and simple-evals graders. On Harbor benchmarks you can switch to a listed harness such as OpenHands, or request your own and compare it. Public-only suites, including Harvey's benchmark and Zapier's AutomationBench, are not directly comparable to the publisher's private held-out board. Arushi Gandhi, co-founder and CEO, posted the launch and says new benchmarks are being added about every six hours, with a request path for ones that are missing. She co-founded the YC Winter 2026 company with CTO Abhishek Eswaran and describes it as New Measure, which is where the product runs.
Tags
Questions
Frequently asked questions
The Ressl launch post reached 145K views on X, with 144 likes, 13 reposts, 46 comments, 21 bookmarks.
Ressl launched on X on November 12, 2026 with the message "Runs public LLM benchmarks on any model or your own harness".
Benchmark Router lets AI builders run public LLM benchmarks on a chosen model, or on an agent harness, without standing up each suite's official runner.
LaunchVideo.com rated the Ressl launch 2.00 out of 5: hook 2/5, editing / creativity 1/5, copy 3/5, sentiment of launch 3/5, distribution strategy 1/5.







