Hugging Face's logo

Datasets:
FINAL-Bench
/
World-Model

Tasks:
Other
Modalities:
Text
Formats:
json
Languages:
English
Korean
Size:
< 1K
Tags:
world-model
embodied-ai
benchmark
agi
cognitive-evaluation
vidraft
Libraries:
Datasets
pandas
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
SeaWolf-AI commited on
Commit
8fdb845
·
verified ·
1 Parent(s): 952db38

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +35 -11
README.md CHANGED
@@ -133,11 +133,45 @@ Then upload `my_submission.json` to the [WM Bench Leaderboard](https://huggingfa
133
 
134
  ---
135
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
136
  ## 📦 Dataset Files
137
 
138
  ```
139
  wm-bench-dataset/
140
- ├── wm_bench_dataset.json # 100 scenarios + ground truth
141
  ├── example_submission.py # Participation template
142
  ├── wm_bench_scoring.py # Scoring engine (fully open)
143
  ├── wm_bench_eval.py # Evaluation runner
@@ -146,16 +180,6 @@ wm-bench-dataset/
146
 
147
  ---
148
 
149
- ## 🏆 Current Leaderboard
150
-
151
- | Rank | Model | Org | WM Score | Grade | Track |
152
- |------|-------|-----|----------|-------|-------|
153
- | 1 | VIDRAFT PROMETHEUS v1.0 | VIDRAFT | 726 | B | C ✓ |
154
-
155
- *Submit your model at the [WM Bench Leaderboard](https://huggingface.co/spaces/FINAL-Bench/worldmodel-bench)*
156
-
157
- ---
158
-
159
  ## 🔬 FINAL Bench Family
160
 
161
  WM Bench is part of the **FINAL Bench Family** — a suite of AGI evaluation benchmarks by VIDRAFT:
 
133
 
134
  ---
135
 
136
+ ## 🏆 Current Leaderboard
137
+
138
+ ![Leaderboard Overview](l1.png)
139
+
140
+ | Rank | Model | Org | WM Score | Grade | Track |
141
+ |------|-------|-----|----------|-------|-------|
142
+ | 1 | VIDRAFT PROMETHEUS v1.0 | VIDRAFT | 726 | B | C ✓ |
143
+
144
+ ![Leaderboard Detail](l2.png)
145
+
146
+ ![Leaderboard Score Breakdown](l3.png)
147
+
148
+ *Submit your model at the [WM Bench Leaderboard](https://huggingface.co/spaces/FINAL-Bench/worldmodel-bench)*
149
+
150
+ ---
151
+
152
+ ## 🌍 PROMETHEUS World Model — Live Demo
153
+
154
+ **WM Bench is powered by VIDRAFT PROMETHEUS**, the world's first real-time embodied AI that combines FloodDiffusion motion generation with a Kimi K2.5 cognitive brain.
155
+
156
+ ![PROMETHEUS World Model](s1.png)
157
+
158
+ > Perceive → Predict → Decide → Act
159
+
160
+ ![PROMETHEUS Scene — Castle World](s2.png)
161
+
162
+ ![PROMETHEUS NPC Interaction](s3.png)
163
+
164
+ ![PROMETHEUS Brain Dashboard](s4.png)
165
+
166
+ 🔗 **Try it live:** [FINAL-Bench/World-Model](https://huggingface.co/spaces/FINAL-Bench/World-Model)
167
+
168
+ ---
169
+
170
  ## 📦 Dataset Files
171
 
172
  ```
173
  wm-bench-dataset/
174
+ ├── wm_bench.jsonl # 100 scenarios + ground truth
175
  ├── example_submission.py # Participation template
176
  ├── wm_bench_scoring.py # Scoring engine (fully open)
177
  ├── wm_bench_eval.py # Evaluation runner
 
180
 
181
  ---
182
 
 
 
 
 
 
 
 
 
 
 
183
  ## 🔬 FINAL Bench Family
184
 
185
  WM Bench is part of the **FINAL Bench Family** — a suite of AGI evaluation benchmarks by VIDRAFT: