List your company

Reward Modeling on EditReward-Bench

85.4PFJRM
40.1651.90563.6575.395Feb 7, 2026Mar 2, 2026Mar 25, 2026Apr 17, 2026May 10, 2026Jun 2, 2026Jun 25, 2026

Evaluation results

MethodLinks
JRM2026.0285.4—85.1
EditReward2026.0283.2—79.2
EditRewardType=Disc.2026.0283.2—79.2
GPT-52026.0277.766.975.5
PaCo-Reward-7BParameters=7B2026.0277.770.975.1
GPT-5Type=Closed, Pointwise VIEScore setting=true2026.0277.766.975.5
GPT-5Type=Closed2026.0677.766.975.5
EditScore-7B (Avg@4)Type=Gen.2026.0672.27272.7
Gemini-3.0-FlashVersion=3.0-Flash2026.0271.766.276.9
Gemini-3.0-FlashType=Closed, Pointwise VIEScore setting=true2026.0271.766.276.9
Gemini-3.0-FlashType=Closed2026.0671.766.276.9
Gemini-2.5-ProVersion=2.5-Pro2026.0270.35672.2
Gemini-2.5-ProType=Closed, Pointwise VIEScore setting=true2026.0270.35672.2
Gemini-2.5-ProType=Closed2026.0670.35672.2
SpatialRewardType=Gen.2026.0268.367.280.3
GPT-4.12026.0267.360.270.5
GPT-4.1Type=Closed, Pointwise VIEScore setting=true2026.0267.360.270.5
GPT-4.1Type=Closed2026.0667.360.270.5
SFReward (Ours)Type=Gen.2026.0664.862.176
EditScore-72BParameters=72B2026.0263.858.670.3
w/o Dense Anno. (Ours)Type=Gen.2026.0662.751.469.9
EditScore-8BParameters=8B2026.0260.859.469
EditScore-8BType=Gen.2026.0260.859.469
Qwen3-VL-8BType=Gen.2026.0241.924.356.2
Qwen3-VL-8BType=Gen.2026.0641.924.356.2