Interaction BenchInteraction Bench
v0.1 preview
LeaderboardEnvironmentsTasksPipelineMethodologyDocsBlogAboutContributors
Github
Grist

Grist

Data & analytics
Spreadsheet that behaves like a relational database
Githubgristlabs/grist-core12.0k
Tech stack
TypeScriptGrainJSNode.jsPythonSQLite
Languages
TypeScript82.9%
Python10.8%
JavaScript5.1%
Other1.2%
Available tools
Browser
Terminal
Network
Keep widget settings aligned after duplicating a page
Keep widget settings aligned after duplicating a page

Tasks

All 1 task
Keep widget settings aligned after duplicating a page
#2298 · Base b675729
Duplicate my document page and continue configuring the selected widget.
33%

Model results

Full leaderboard
Model
Harness
DeepSeek
DeepSeek Flash
DSH
15.8M in · 43.2k out · 15.8M total
DSH · 154 steps · $0.09 · 15m 47s · 15.8M tokens
DSH
154
$0.09
15m 47s
15.8M
43.2k
15.8M
OpenAI
GPT-6 Astra
Codex
3.0M in · 10.8k out · 3.0M total
Codex · 49 steps · $4.26 · 9m 59s · 3.0M tokens
Codex
Codex
49
$4.26
9m 59s
3.0M
10.8k
3.0M
MoonshotAI
Kimi K3
Kimi Code
3.4M in · 25.2k out · 3.5M total
Kimi Code · 62 steps · $1.60 · 20m 37s · 3.5M tokens
Kimi
Kimi Code
62
$1.60
20m 37s
3.4M
25.2k
3.5M
Z.ai
GLM-5.3 Flash
ZCode
71.1M in · 136.5k out · 71.3M total
ZCode · 389 steps · $2.26 · 1h 03m · 71.3M tokens
ZCode
389
$2.26
1h 03m
71.1M
136.5k
71.3M
Claude
Claude Opus 5.5
Claude Code
6.5M in · 42.9k out · 6.6M total
Claude Code · 85 steps · $3.10 · 29m 44s · 6.6M tokens
Claude Code
Claude Code
85
$3.10
29m 44s
6.5M
42.9k
6.6M
Qwen
Qwen3.8 Max
Codex
36.6M in · 84.6k out · 36.7M total
Codex · 287 steps · $10.50 · 1h 03m · 36.7M tokens
Codex
Codex
287
$10.50
1h 03m
36.6M
84.6k
36.7M
Interaction BenchInteraction Bench

Benchmarked on what users feel.

// results

LeaderboardEnvironmentsTasks

// how it works

PipelineMethodologyDocsBlog

// project

AboutContributorsContact usGitHubReport an issue
$ ibench --release 0.1 --statuspreview · built by EvalSys