List your company

Mathematical Reasoning on GSM-sym

90.22Exact MatchR1-Oracle
2.610425.355248.170.8448Sep 5, 2025Oct 6, 2025Nov 7, 2025Dec 8, 2025Jan 9, 2026Feb 9, 2026Mar 13, 2026

Evaluation results

MethodLinks
R1-OracleBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=100% GT2026.0390.2279.120—
SA-GRPOBase Model=Qwen3-4B, alpha=0.22025.0987.64——323
GRPO-O1Base Model=Qwen3-4B, alpha=0.052025.0986.76——350
MemRewardBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=20% GT + GNN2026.0386.4477.02-2.1—
GRPOBase Model=Qwen3-4B2025.0986.3——414
PPOBase Model=Qwen3-4B2025.0985.42——485
GRPO-ERBase Model=Qwen3-4B, alpha=0.052025.0985.24——297
R1-pBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=20% GT2026.0384.6775.67-3.45—
SA-PPOBase Model=Qwen3-4B2025.0983.52——345
SA-GRPOBackbone=Gemma3-4B2025.0979.04——654
GRPOBackbone=Gemma3-4B2025.0978.68——644
MemRewardBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=20% GT + GNN2026.0377.7868.1-2.37—
R1-OracleBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=100% GT2026.0375.3370.470—
GRPO-O1Backbone=Gemma3-4B2025.0975.02——464
GRPOBase Model=Phi-3.5-mini2025.0973.76——337
SA-PPOBase Model=Phi-3.5-mini2025.0973.52——344
GRPO-ERBackbone=Gemma3-4B2025.0973.36——623
SA-GRPOBase Model=Phi-3.5-mini, alpha=0.22025.0972.72——289
SA-PPOBase Model=Qwen3-1.7B2025.0968.02——369
GRPOBase Model=Qwen3-1.7B2025.0967.66——413
SA-GRPOBase Model=Qwen3-1.7B, alpha=0.22025.0967.66——346
Qwen3-4BBase Model=Qwen3-4B2025.0966.22——435
GRPO-O1Base Model=Qwen3-1.7B, alpha=0.052025.0965.78——363
PPOBase Model=Phi-3.5-mini2025.0964.88——322
GRPO-ERBase Model=Qwen3-1.7B, alpha=0.052025.0964.74——320
GRPO-ERBase Model=Phi-3.5-mini, alpha=0.052025.0963.54——240
R1-pBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=20% GT2026.0362.8962.72-7.75—
PPOBase Model=Qwen3-1.7B2025.0962.88——460
GRPO-O1Base Model=Phi-3.5-mini, alpha=0.052025.0962.24——279
Gemma3-4BBackbone=Gemma3-4B, Strategy=Base2025.0957.4——373
Phi-3.5-miniBase Model=Phi-3.5-mini2025.0956.86——398
Qwen3-1.7BBase Model=Qwen3-1.7B2025.0946.76——365
GRPOBase Model=Gemma3-1B2025.0934.7——1,552
SA-GRPOBase Model=Gemma3-1B, alpha=0.22025.0934.52——772
GRPO-ERBase Model=Gemma3-1B, alpha=0.052025.0932.36——1,063
GRPO-O1Base Model=Gemma3-1B, alpha=0.052025.0932.06——1,157
SA-PPOBase Model=Gemma3-1B2025.0931——936
PPOBase Model=Gemma3-1B2025.0930.28——1,187
GRPOBackbone=Qwen2-2B-VL2025.0927.72——209
SA-GRPOBackbone=Qwen2-2B-VL2025.0927.66——174
GRPO-ERBackbone=Qwen2-2B-VL2025.0926.52——179
GRPO-O1Backbone=Qwen2-2B-VL2025.0925.44——164
Gemma3-1BBase Model=Gemma3-1B2025.0919.58——427
Qwen2-2B-VLBackbone=Qwen2-2B-VL, Strategy=Base2025.095.98——1,873