List your company
ResearchFormerly WizWand

Mathematics on AIME 25 (Score (%))

61.11Score (%)Qwen3.5-9B + AR-SFT
5.771620.138334.50548.8717Apr 17, 2026Apr 24, 2026May 2, 2026May 9, 2026May 17, 2026May 24, 2026Jun 1, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Qwen3.5-9B + AR-SFTModel Size=9B, Training Protocol=AR-SFT, Decoding Strategy=Native AR decoding2026.0661.11
Qwen3.5-9B (released)Model Size=9B, Training Protocol=Released, Decoding Strategy=Native AR decoding2026.0660
FLARE-9BModel Size=9B, Training Protocol=FLARE, Decoding Strategy=AR-Trust sampling2026.0654.44
Qwen3.5-4B (released)Model Size=4B, Training Protocol=Released, Decoding Strategy=Native AR decoding2026.0648.89
Qwen3.5-4B + AR-SFTModel Size=4B, Training Protocol=AR-SFT, Decoding Strategy=Native AR decoding2026.0646.67
FLARE-4BModel Size=4B, Training Protocol=FLARE, Decoding Strategy=AR-Trust sampling2026.0643.33
MPO+Backbone=olmo-3.1-32b-it-sft2026.0435.8
Mag+Backbone=olmo-3.1-32b-it-sft2026.0433.3
Soft+Backbone=olmo-3.1-32b-it-sft2026.0432.3
Pair+Backbone=olmo-3.1-32b-it-sft2026.0431.7
Mag+Backbone=olmo-3-7b-it-sft2026.0428.1
RFTBackbone=olmo-3.1-32b-it-sft2026.0427.5
Pair+Backbone=olmo-3-7b-it-sft2026.0426.9
FLARE-2BModel Size=2B, Training Protocol=FLARE, Decoding Strategy=AR-Trust sampling2026.0626.67
Soft+Backbone=olmo-3-7b-it-sft2026.0426.5
MPO+Backbone=olmo-3-7b-it-sft2026.0426
DPOBackbone=olmo-3.1-32b-it-sft2026.0425.6
DPOBackbone=olmo-3-7b-it-sft2026.0425.2
Qwen3.5-2B + AR-SFTModel Size=2B, Training Protocol=AR-SFT, Decoding Strategy=Native AR decoding2026.0622.22
RFTBackbone=olmo-3-7b-it-sft2026.0417.3
Qwen3.5-2B (released)Model Size=2B, Training Protocol=Released, Decoding Strategy=Native AR decoding2026.0612.22
BaseBackbone=olmo-3.1-32b-it-sft2026.048.1
BaseBackbone=olmo-3-7b-it-sft2026.047.9