List your company

Reasoning on ARC Challenge (accuracy, length)

418Solution Length+ShorterBetter
378.8643.49081,172.6May 8, 2026

Evaluation results

MethodLinks
+ShorterBetterBackbone=Qwen3-4B2026.054180.9297
+LP-G (Kimi)Backbone=Qwen3-4B2026.054640.923
+LC-R1Backbone=Qwen3-4B2026.055050.9163
+ICR /w LP-FBackbone=Qwen3-8B2026.056020.9397
+ICR /w LP-FBackbone=DSQW-7B2026.056030.8428
+ICR /w LP-FBackbone=Qwen3-4B2026.056080.9397
+ShorterBetterBackbone=Qwen3-8B2026.056140.9197
+ICRBackbone=Qwen3-4B2026.056220.9364
+LP-G (Kimi)Backbone=Qwen3-8B2026.056460.9264
+LP-FBackbone=Qwen3-8B2026.056990.9364
+LC-R1Backbone=Qwen3-8B2026.057010.9264
+ShorterBetterBackbone=DSQW-7B2026.057070.8127
+LP-FBackbone=Qwen3-4B2026.057320.9364
+LP-FBackbone=DSQW-7B2026.057920.826
+ICRBackbone=Qwen3-8B2026.058040.9331
+LP-G (Kimi)Backbone=DSQW-7B2026.058550.796
+LC-R1Backbone=DSQW-7B2026.058780.8327
+GRPOBackbone=Qwen3-4B2026.058860.9257
+GRPOBackbone=Qwen3-8B2026.059810.9257
+ICRBackbone=DSQW-7B2026.051,1770.8261
BaseBackbone=Qwen3-4B2026.051,2310.9331
BaseBackbone=DSQW-7B2026.051,3210.786
+GRPOBackbone=DSQW-7B2026.051,3860.816
BaseBackbone=Qwen3-8B2026.051,3980.9264