Share your thoughts, 1 month free Claude Pro on usSee more

Multi-hop Reasoning on RULER QA

95.4Accuracy (32K Context)

Qwen2.5-14B-1M-LongRLVR

Updated 1mo ago

Evaluation Results

Method	Links
Qwen2.5-14B-1M-LongRLVR 2026.03		95.4	87.8	83.5	88.9
Qwen3-14B (Thinking) 2026.03		91.2	89	82.6	87.6
Qwen2.5-14B-1M 2026.03		90.6	70.6	64.4	75.2
QwenLong-L1-32B 2026.03		89	77	72.4	79.47
Qwen2.5-14B-1M-SFT 2026.03		88	66.5	62.2	72.23
Qwen3-8B (Thinking) 2026.03		86.5	84	81.8	84.1
Qwen2.5-14B-1M-RLVR 2026.03		86.3	69	64.2	73.17
LLaMA-3.1-8B-LongRLVR 2026.03		85.5	76.5	79	80.33
Qwen2.5-7B-1M-LongRLVR 2026.03		82.5	76.5	77	78.67
Qwen2.5-7B-1M-RLVR 2026.03		74.4	68.5	57.8	66.9
Qwen2.5-7B-1M-SFT 2026.03		72.4	64.2	56.8	64.47
LLaMA-3.1-8B-RLVR 2026.03		72	68.8	62.6	67.8
Qwen2.5-7B-1M 2026.03		70.5	66	58.5	65
LLaMA-3.1-70B 2026.03		70.4	64.2	47.6	60.73
LLaMA-3.1-8B-SFT 2026.03		68.4	65.3	60.4	64.7
Qwen2.5-72B-YaRN 2026.03		66.9	54.5	47.2	56.2
LLaMA-3.1-8B 2026.03		65.8	63.7	58.8	62.77