Downloads · 30 days
0
verantyx-arc-agi2
kofdai/verantyx-arc-agi2
verantyx-arc-agi2 is a machine learning model from kofdai. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
Pure rule-based solver for ARC-AGI-2. No LLM. No neural networks. No pre-training.
Access
Public
Updated Feb 24, 2026
Repo size
—
Likes
1
Public
Repo makeup
Click a slice to open those files.
.py693 KB · 99%
At a glance
- License
- mit
- Access
- Public
- Created
- Feb 24, 2026
- Updated
- Feb 24, 2026
- SHA
- 7eddc0fc
Datasets
Model card
From the Hugging Face model README
Verantyx ARC-AGI-2 — 16.1% (161/1000) 🏆
Pure rule-based solver for ARC-AGI-2. No LLM. No neural networks. No pre-training.
Score
| Split | Score | Tasks Solved |
|---|---|---|
| Training | 16.1% | 161 / 1000 |
Approach
Verantyx is a fully deterministic, LLM-free ARC-AGI-2 solver built on program synthesis and domain-specific language (DSL) enumeration.
Architecture
- Cross Engine — Multi-strategy solver with priority-based candidate selection
- Neighborhood Rules — Exact and extended (count/directional/multi-pass) neighborhood matching
- DSL Enumerator — 32 parameterless primitives with depth-2 composition (1024 combinations)
- Panel Operations — Grid splitting (separators, equal partitions) + reduction (XOR, OR, AND, overlay, select)
- Per-Object Transform — Object detection → property-based recoloring/movement
- Puzzle Language — 25+ hand-crafted pattern detectors with auto-verification
- Iterative Cross — 2-step residual learning for complex transformations
Puzzle Language Patterns (Phase 7)
| Pattern | Description | Example Task |
|---|---|---|
cross_and_fill | Fill cross intersections | — |
diagonal_extend | Extend diagonal lines | — |
connect_same_color | Connect same-colored cells | — |
extract_tile | Detect repeated tile | 7b7f7511 |
two_row_interleave | 2 rows → checkerboard | e9afcf9a |
stack_nor | Top+bottom halves NOR | fafffa47 |
frame_repeat_border | Frame with repeated border | 49d1d64f |
col_color_map | Column position → row color | a85d4709 |
split_vsep_and | Vertical split AND | 0520fde7 |
grid_pattern | Lattice/checkerboard generation | 332efdb3 |
antidiag_fill | Anti-diagonal + bottom fill | 3bd67248 |
latin_square | Latin square completion | 4cd1b7b2 |
shift_recolor | Shift + recolor fg cells | a79310a0 |
uniform_row_detect | Mark uniform rows | 25d8a9c8 |
| + 10 more | ... | ... |
Score History
| Version | Score | Delta | Key Changes |
|---|---|---|---|
| v19 | 113 (11.3%) | — | Baseline with NB + DSL |
| v27 | 127 (12.7%) | +14 | Extended NB rules |
| v28 | 136 (13.6%) | +9 | split_combine + panel_extract |
| v29 | 142 (14.2%) | +6 | puzzle_lang Phase 7 |
| v30 | 144 (14.4%) | +2 | connect_same_color |
| v31 | 146 (14.6%) | +2 | u_drop_ball + nonzero_count_row |
| v32 | 147 (14.7%) | +1 | scale_down_uniform |
| v33 | 149 (14.9%) | +2 | staircase_grow + comp_tile_2x2 |
| v34 | 154 (15.4%) | +5 | frame_repeat + extract_tile |
| v35 | 158 (15.8%) | +4 | col_color_map + split_vsep_and |
| v36 | 161 (16.1%) | +3 | grid_pattern + latin_square |
Performance
- Eval speed: ~0.40s/task (1000 tasks in ~7 minutes)
- No GPU required — runs on CPU only
- Deterministic — same input always produces same output
Usage
from arc.cross_engine import solve_cross_engine
# train_pairs: list of (input_grid, output_grid)
# test_inputs: list of input grids
predictions, versions = solve_cross_engine(train_pairs, test_inputs)
Files
arc/cross_engine.py— Main solver orchestratorarc/puzzle_lang.py— Pattern-based puzzle languagearc/nb_extended.py— Extended neighborhood rulesarc/beam_search.py— Beam search with DSLarc/enumerator.py— DSL primitive enumeratorarc/panel_ops.py— Panel split + reduce operationsarc/per_object.py— Per-object transformsarc/primitives.py— 32 core DSL primitives
License
MIT
Author
kofdai — verantyx project
- License
- mit
- Languages
- en
- Created
- Feb 24, 2026
- Updated
- Feb 24, 2026