List your company
ResearchFormerly WizWand

Competitive Programming on LiveCodeBench v5 (test)

61.6ScoreKlear-Reasoner
14.38426.64238.951.158Jan 11, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Klear-ReasonerBase Model=Klear-Reasoner-SFT, SFT=true, RL=true, Model Size=8B, Training Data Volume=106k, Training Task Type=Real, Evaluation Protocol=avg@82026.0161.6
Klear-Reasoner-SFTBase Model=Qwen3-Base, SFT=true, RL=false, Model Size=8B, Training Data Volume=1500k, Training Task Type=Real, Evaluation Protocol=avg@82026.0158.5
AReal-boba²Base Model=R1-Distilled-Qwen, SFT=false, RL=true, Model Size=14B, Training Data Volume=24k, Training Task Type=Real, Evaluation Protocol=avg@322026.0158.1
DeepCoder-PreviewBase Model=R1-Distilled-Qwen, SFT=false, RL=true, Model Size=14B, Training Data Volume=24k, Training Task Type=Real, Evaluation Protocol=pass@12026.0157.9
MiMoBase Model=MiMo-SFT, SFT=true, RL=true, Model Size=7B, Training Data Volume=130k, Training Task Type=Undisclosed, Evaluation Protocol=avg@82026.0157.8
Qwen3-8BBase Model=Qwen3-8B-Base, SFT=true, RL=false, Model Size=8B, Training Task Type=Real2026.0157.5
rStar-CoderBase Model=Qwen2.5-Coder-Instruct, SFT=true, RL=false, Model Size=7B, Training Data Volume=580K, Training Task Type=Mixed, Evaluation Protocol=avg@162026.0157.3
AceReason1.1Base Model=AceReason1.1-SFT, SFT=true, RL=true, Model Size=7B, Training Task Type=Real, Evaluation Protocol=avg@82026.0157.2
MiMo-SFTBase Model=MiMo-Base, SFT=true, RL=false, Model Size=7B, Training Data Volume=500k, Training Task Type=Undisclosed, Evaluation Protocol=avg@82026.0152.3
OpenThinker3Base Model=Qwen2.5-Instruct, SFT=true, RL=false, Model Size=7B, Training Data Volume=1,200k, Training Task Type=Mixed2026.0151.7
OCR-Qwen-InstructBase Model=Qwen2.5-Instruct, SFT=true, RL=false, Model Size=7B, Training Data Volume=736k, Training Task Type=Real, Evaluation Protocol=avg@642026.0151.3
AceReason1.1-SFTBase Model=Qwen2.5-Math, SFT=true, RL=false, Model Size=7B, Training Data Volume=2.2M, Training Task Type=Real, Evaluation Protocol=avg@82026.0151.2
Skywork-OR1Base Model=R1-Distilled-Qwen, SFT=false, RL=true, Model Size=7B, Training Data Volume=124k, Training Task Type=Real, Evaluation Protocol=avg@322026.0147.6
OlympicCoderBase Model=Qwen2.5-Coder-Instruct, SFT=true, RL=false, Model Size=7B, Training Data Volume=100k, Training Task Type=Real2026.0140.9
Bespoke-StratosBase Model=Qwen2.5-Instruct, SFT=true, RL=false, Model Size=7B, Training Data Volume=17k, Training Task Type=Real, Evaluation Protocol=pass@12026.0116.2