OpenCodePapers

agents-on-pinchbench

Agents
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeAccuracyModelNameReleaseDate
KAT-Coder-V2.5 Technical Report94.9KAT-Coder-V2.52026-07-06
Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning✓ Link90.0Nemotron 3 Ultra2026-06-03
KAT-Coder-V2 Technical Report88.7KAT-Coder-V22026-03-29
Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale✓ Link87.60Ring-2.6-1T (high)2026-06-13
Ling-2.6-1T: A Trillion-Parameter Comprehensive Flagship Model for Complex Tasks87.40Ling-2.6-1T2026-04-29
ClawGym: A Scalable Framework for Building Effective Claw Agents86.0ClawGym-30A3B2026-04-29
NVIDIA Nemotron 3.5 Lightning✓ Link85.37NVIDIA Nemotron 3.5 Lightning 30B-A3B2026-08-11
Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale✓ Link85.24Ling-2.6-1T2026-06-13
LFM2.5-24B-A2B: Built for Fast On-Device Agents83.18LFM2.5-24B-A2B2026-10-13
NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness✓ Link82.25NeoHorse-1-9B2026-09-08
Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale✓ Link81.30Ling-2.6-flash2026-06-13
Mach-Mind-4-Flash Technical Report74.90Mach-Mind-4-Flash2026-07-10
LFM2.5-2.6B: Deploy Agents Everywhere✓ Link68.22LFM2.5-2.6B2026-08-04
Check out how to contribute new results. Then edit this file on GitLab or GitHub.