Benchmarks

This page documents Rezi’s benchmark methodology, results, and comparisons with other TUI frameworks. Source: BENCHMARKS.md in repository
Suite: packages/bench/
Last updated: 2026-02-22

Frameworks Compared

Framework	Runtime	Driver	Notes
Rezi	Node.js (worker)	`@rezi-ui/core` + `@rezi-ui/node`	Full layout engine, binary drawlist, native C renderer
Ink	Node.js (worker)	React + Yoga + ANSI output	React reconciler, flexbox layout via Yoga
OpenTUI (React)	Bun (subprocess)	`@opentui/react` on `@opentui/core`	React declarative driver
OpenTUI (Core)	Bun (subprocess)	`@opentui/core` imperative API	Direct object mutation, no React overhead
Bubble Tea	Go (subprocess)	`charmbracelet/bubbletea`	Elm architecture, string-based `View()` rendering
terminal-kit	Node.js (worker)	`terminal-kit`	Low-level terminal buffer library, no widget/layout system
blessed	Node.js (worker)	`blessed`	Low-level terminal UI with box model, no constraint layout
Ratatui	Rust (subprocess)	`ratatui`	Native Rust terminal renderer, immediate mode

Benchmark Scenarios

The suite contains 22 scenarios in three tiers:

Primitive Workloads

Isolated subsystem stress tests:

startup — app create/destroy cycle
tree-construction — build widget tree from scratch (10, 100, 500, 1000 items)
rerender — update a single changing value in an existing tree
content-update — update text content across a large tree
layout-stress — grid layout with many cells
scroll-stress — scroll a 2000-item list
virtual-list — viewport-windowed rendering over 100K items
tables — 100-row, 8-column data table rendering
memory-profile — steady-state memory behavior

Terminal-Level Workloads

End-to-end render + terminal write:

terminal-rerender
terminal-frame-fill (1 and 40 dirty lines)
terminal-screen-transition
terminal-fps-stream
terminal-input-latency
terminal-memory-soak
terminal-virtual-list
terminal-table

Full-App Workloads

Structured UI composition:

terminal-full-ui — composite dashboard shell with panels, status bar, data sections
terminal-full-ui-navigation — same shell with page routing
terminal-strict-ui — structured multi-panel layout (header, 3-column body, footer, status bar)
terminal-strict-ui-navigation — same structured layout with navigation

Strict scenarios exist to reduce bias when comparing Rezi (which has a layout engine) against libraries that operate closer to raw terminal buffers (blessed, ratatui, terminal-kit).

Results (2026-02-22, PTY mode)

Single-replicate, default iteration counts. WSL host (directional only).

Rezi vs. Ink

21 scenarios where both participate:

Scenario	Ratio (Ink slower)
tree-construction (10 items)	206x
tree-construction (100 items)	80x
tree-construction (500 items)	49x
tree-construction (1000 items)	46x
rerender	47x
content-update	32x
layout-stress	10x
scroll-stress	12x
virtual-list	23x
tables	11x
terminal-rerender	47x
terminal-frame-fill (1 dirty line)	55x
terminal-frame-fill (40 dirty lines)	36x
terminal-fps-stream	10x
terminal-input-latency	35x
terminal-full-ui	10x
terminal-strict-ui	21x

Summary: Rezi is faster than Ink in every measured scenario. The gap is largest on tree construction and rerender workloads (45-206x) and smallest on complex UI composition (10x).

Rezi vs. OpenTUI (React)

21 scenarios: Rezi faster in all 21 scenarios. Range: 1.8x to 155x. Geomean: ~10x.

Rezi vs. OpenTUI (Core)

21 scenarios: Rezi faster in 19/21 scenarios. Range: 1.3x to 13x. Geomean: ~2.6x. OpenTUI Core wins:

layout-stress: Core 1.5x faster (517 vs 347 ops/s)
tables: Core 1.6x faster (434 vs 277 ops/s)

Rezi vs. Bubble Tea

21 scenarios: Rezi faster in 20/21 scenarios. Bubble Tea throughput clusters around ~120 ops/s in most scenarios due to its 8.33ms tick rate. Bubble Tea wins:

scroll-stress: Bubble Tea 2.5x faster (120 vs 48 ops/s)

Rezi vs. terminal-kit / blessed / Ratatui

These three participate only in scenarios that make sense for their level of abstraction. On primitive workloads (rerender, frame-fill, simple construction):

terminal-kit: 1x-13x faster than Rezi
blessed: 2x-19x faster than Rezi
Ratatui: 2x-20x faster than Rezi

On complex UI workloads (strict-ui, full-ui, fps-stream, tables, virtual-list):

Rezi is competitive or faster

Rationale: terminal-kit and blessed write directly to terminal buffers without computing layout. Ratatui is compiled Rust. On workloads that actually require structured layout and widget composition, the gap narrows or reverses.

Representative Numbers

Selected scenarios showing the performance landscape across categories:

Scenario	Rezi	Ink	OpenTUI React	OpenTUI Core	Bubble Tea	Ratatui
startup	1.87ms (516 ops/s)	5.62ms (112)	8.68ms (33)	4.92ms (38)	9.94ms (49)	184µs (5.2K)
tree-construction (100)	326µs (3.1K)	26ms (38)	36ms (27)	2.15ms (466)	8.33ms (120)	696µs (1.4K)
rerender	373µs (2.7K)	17.7ms (57)	2.70ms (370)	1.16ms (860)	8.33ms (120)	51µs (19.7K)
layout-stress	2.88ms (347)	28ms (36)	33ms (30)	1.93ms (517)	8.33ms (120)	—
virtual-list (100K items)	985µs (1.0K)	22.6ms (44)	28.5ms (35)	1.28ms (780)	8.33ms (120)	—
terminal-strict-ui	1.19ms (836)	25.5ms (39)	19.4ms (51)	1.77ms (565)	8.33ms (120)	240µs (4.2K)
terminal-full-ui	2.49ms (401)	25.6ms (39)	5.07ms (197)	1.31ms (760)	8.33ms (120)	336µs (3.0K)

Memory Usage

Framework	Typical peak RSS (UI scenarios)	Notes
Rezi	80-210 MB	Heap ~20-120 MB depending on tree size
Ink	120-980 MB	Grows significantly with tree size
OpenTUI (React)	200 MB - 15 GB	Memory scales poorly; OOMs on tree-construction at 1000 items
OpenTUI (Core)	100-190 MB	Comparable to Rezi
Bubble Tea	7-10 MB	Go runtime baseline, very low footprint
Ratatui	3-16 MB	Native binary, minimal overhead
terminal-kit	69-83 MB	Lightweight buffer library
blessed	77-300 MB	Varies with screen complexity

Methodology

What is Measured

Each benchmark iteration measures one complete render cycle: state update through final frame output. For PTY-mode benchmarks, this includes the terminal write path.

Runtime Differences

These benchmarks compare frameworks, but also inevitably compare runtimes:

Rezi, Ink, blessed: Node.js worker processes
OpenTUI (both drivers): Bun subprocesses
Bubble Tea: Go subprocess
Ratatui: Rust subprocess

Cross-framework numbers include runtime and toolchain effects alongside framework costs.

Scenario Parity

Primitive scenarios (rerender, frame-fill) test narrow subsystem performance. Libraries without layout engines (terminal-kit, blessed, Ratatui) have an inherent advantage here because they do less work per frame.
Strict-ui scenarios build equivalent multi-panel layouts across all frameworks. Bubble Tea renders via lipgloss string composition rather than a widget tree, which is a different code path.
Full-ui scenarios test composite dashboards with data sections, status bars, and navigation.

Environment

WSL and virtualized hosts introduce measurable scheduler jitter. Treat WSL-collected results as directional. For publication-grade numbers, use bare-metal Linux with CPU pinning and multiple replicates.

Reproducing Benchmarks

Full Cross-Framework Suite (PTY mode)

npm ci
npm run build
npm run build:native
npx tsc -b packages/bench

# Build Ratatui bench binary
cd benchmarks/native/ratatui-bench
cargo build --release
cd -

node --expose-gc packages/bench/dist/run.js \
  --suite all \
  --io pty \
  --output-dir benchmarks/local-all

Rigorous Terminal Suite (replicates + shuffle + affinity)

node --expose-gc packages/bench/dist/run.js \
  --suite terminal \
  --io pty \
  --replicates 7 \
  --discard-first-replicate \
  --shuffle-framework-order \
  --shuffle-seed local-terminal \
  --cpu-affinity 0-7 \
  --env-check warn \
  --output-dir benchmarks/local-terminal

Quick Matchup: Rezi vs OpenTUI

# React driver
node --expose-gc packages/bench/dist/run.js \
  --matchup rezi-opentui \
  --io pty \
  --quick \
  --output-dir benchmarks/local-rezi-opentui-react-quick

# Core driver
node --expose-gc packages/bench/dist/run.js \
  --matchup rezi-opentui \
  --opentui-driver core \
  --io pty \
  --quick \
  --output-dir benchmarks/local-rezi-opentui-core-quick

Limitations

PTY benchmarks include framework render + terminal write path, not terminal emulator pixel paint.
Absolute numbers are host-specific; compare within the same dataset, mode, and host.
Quick-mode single-replicate outputs are useful for trend checks, not confidence-grade claims.
OpenTUI React OOMs on tree-construction at 1000 items (inherent to React driver memory scaling).

Render Pipeline — Rendering phases
Layout Engine — Layout algorithm details
Determinism — Reproducibility guarantees

Overview

Protocol

Backend

Reference

Frameworks Compared

Benchmark Scenarios

Primitive Workloads

Terminal-Level Workloads

Full-App Workloads

Results (2026-02-22, PTY mode)

Rezi vs. Ink

Rezi vs. OpenTUI (React)

Rezi vs. OpenTUI (Core)

Rezi vs. Bubble Tea

Rezi vs. terminal-kit / blessed / Ratatui

Representative Numbers

Memory Usage

Methodology

What is Measured

Runtime Differences

Scenario Parity

Environment

Reproducing Benchmarks

Full Cross-Framework Suite (PTY mode)

Rigorous Terminal Suite (replicates + shuffle + affinity)

Quick Matchup: Rezi vs OpenTUI

Limitations

Build docs developers (and LLMs) love

Overview

Protocol

Backend

Reference

​Frameworks Compared

​Benchmark Scenarios

​Primitive Workloads

​Terminal-Level Workloads

​Full-App Workloads

​Results (2026-02-22, PTY mode)

​Rezi vs. Ink

​Rezi vs. OpenTUI (React)

​Rezi vs. OpenTUI (Core)

​Rezi vs. Bubble Tea

​Rezi vs. terminal-kit / blessed / Ratatui

​Representative Numbers

​Memory Usage

​Methodology

​What is Measured

​Runtime Differences

​Scenario Parity

​Environment

​Reproducing Benchmarks

​Full Cross-Framework Suite (PTY mode)

​Rigorous Terminal Suite (replicates + shuffle + affinity)

​Quick Matchup: Rezi vs OpenTUI

​Limitations

​Related Documentation

Build docs developers (and LLMs) love

Frameworks Compared

Benchmark Scenarios

Primitive Workloads

Terminal-Level Workloads

Full-App Workloads

Results (2026-02-22, PTY mode)

Rezi vs. Ink

Rezi vs. OpenTUI (React)

Rezi vs. OpenTUI (Core)

Rezi vs. Bubble Tea

Rezi vs. terminal-kit / blessed / Ratatui

Representative Numbers

Memory Usage

Methodology

What is Measured

Runtime Differences

Scenario Parity

Environment

Reproducing Benchmarks

Full Cross-Framework Suite (PTY mode)

Rigorous Terminal Suite (replicates + shuffle + affinity)

Quick Matchup: Rezi vs OpenTUI

Limitations

Related Documentation