Preflight by Grayworth
Connecting Gray language
Gray runtime intelligence

Benchmark Index

Measured performance across Gray execution layers, release candidates, and controlled hosts.

Source release
Latest retained focused comparison · measured August 16, 2026

Resident strict typed JSON: Gray Static vs Rust

Two retained runs each contain 61 paired rounds of strict flat JSON decoding plus checksum. Net duration subtracts the matched control; lower is better. Rust 1.96.0 with serde_json, Linux x86_64 on AMD EPYC 9645 (KVM).

Median net duration per measured body, 2,621,440 decodes; ratio is Gray / Rust
RunGray StaticRustGray duration95% ratio interval
Run 11,282.650 ms1,523.771 ms15.82% lower0.8130–0.8638
Run 21,260.596 ms1,502.832 ms16.12% lower0.8221–0.8651

Source: 7072d8cf28be4d127e313a53873e4c2238eeec71. These are retained measurements, not a fresh rerun or a general language ranking. They do not qualify the distributed RC16 build or the release selected below.

View raw net observations, packet hashes, and measurement scope (JSON)

Selected candidate - Reading release state
Runtime matrix - Checking release coverage
Current comparisons - Fixture and host scoped
Qualification - Automation and runtime gates
Release coverage Checking runtime matrix

Reading standard and sustained benchmark slices.

Comparable evidence

Runtime leaderboards

Rankings never cross fixtures, hosts, metrics, or declared comparison scopes.

Loading benchmark Preparing comparable results Whole-process duration. Execution policies and controls are shown separately.
Rank Implementation Variant Result vs leader Trials Environment
Evidence landscape

Current Gray results by layer

Accepted Gray measurements only. Values remain on their original fixture scales.