Interaction BenchInteraction Bench
v0.1 preview
LeaderboardEnvironmentsTasksPipelineMethodologyDocsBlogAboutContributors
Github
Dify

Dify

AI platform
Builder for agentic workflows and RAG apps
Githublanggenius/dify159.0k
Tech stack
Next.jsReactFlaskPythonPostgreSQL
Languages
Python54.6%
TypeScript41.9%
JavaScript2%
Other1.5%
Available tools
Browser
Terminal
Network
Use Features and Publish in Chatflow Studio after a publish toast closes
Use Features and Publish in Chatflow Studio after a publish toast closes

Tasks

All 1 task
Use Features and Publish in Chatflow Studio after a publish toast closes
#38063 · Base 8218694
Use Features and Publish in Chatflow Studio.
0%

Model results

Full leaderboard
Model
Harness
DeepSeek
DeepSeek Flash
DSH
32.2M in · 55.9k out · 32.3M total
DSH · 160 steps · $0.35 · 26m 25s · 32.3M tokens
DSH
160
$0.35
26m 25s
32.2M
55.9k
32.3M
Z.ai
GLM-5.3 Flash
ZCode
19.7M in · 66.2k out · 19.7M total
ZCode · 207 steps · $0.65 · 1h 06m · 19.7M tokens
ZCode
207
$0.65
1h 06m
19.7M
66.2k
19.7M
Claude
Claude Opus 5.5
Claude Code
7.0M in · 25.6k out · 7.0M total
Claude Code · 80 steps · $2.51 · 2h 01m · 7.0M tokens
Claude Code
Claude Code
80
$2.51
2h 01m
7.0M
25.6k
7.0M
Qwen
Qwen3.8 Max
Codex
16.8M in · 49.4k out · 16.8M total
Codex · 170 steps · $5.11 · 1h 09m · 16.8M tokens
Codex
Codex
170
$5.11
1h 09m
16.8M
49.4k
16.8M
OpenAI
GPT-6 Astra
Codex
4.1M in · 10.3k out · 4.1M total
Codex · 55 steps · $5.56 · 21m 24s · 4.1M tokens
Codex
Codex
55
$5.56
21m 24s
4.1M
10.3k
4.1M
MoonshotAI
Kimi K3
Kimi Code
26.5M in · 144.5k out · 26.7M total
Kimi Code · 120 steps · $11.85 · 1h 06m · 26.7M tokens
Kimi
Kimi Code
120
$11.85
1h 06m
26.5M
144.5k
26.7M
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