List your company

Scientific Reasoning on SuperGPQA (Accuracy)

44.7AccuracyKimi-K2 Base
11.10819.82928.5537.271Jan 6, 2026Jan 8, 2026Jan 10, 2026Jan 13, 2026Jan 15, 2026Jan 17, 2026Jan 20, 2026

Evaluation results

MethodLinks
Kimi-K2 Base# Shots=5-shot, # Activated Params=32B, # Total Params=1043B2026.0144.7
DeepSeek-V3.2 Exp Base# Shots=5-shot, # Activated Params=37B, # Total Params=671B2026.0143.6
DeepSeek-V3.1 Base# Shots=5-shot, # Activated Params=37B, # Total Params=671B2026.0142.3
MiMo-V2-Flash Base# Shots=5-shot, # Activated Params=15B, # Total Params=309B2026.0141.1
Jet-RLModel=Qwen3-8B-Base, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0135.2
BF16Model=Qwen3-8B-Base, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0135.1
InitialModel=Qwen3-8B-Base, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0131.8
BF16-Train-FP8-RolloutModel=Qwen3-8B-Base, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0130.3
BF16Model=Qwen2.5-7B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0128.6
Jet-RLModel=Qwen2.5-7B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0128.5
InitialModel=Qwen2.5-7B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0125.5
Jet-RLModel=Llama3.1-8B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0119.9
BF16Model=Llama3.1-8B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0115.9
BF16-Train-FP8-RolloutModel=Llama3.1-8B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0114.7
InitialModel=Llama3.1-8B, Rollout Length=8k, Training Dataset (RL)=GSM8k + MATH2026.0112.4