| DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | ✓ Link | 93.5 | DeepSeek-V4-Pro Max | 2026-04-26 |
| Solar Open 2 Technical Report | ✓ Link | 92.4 | Solar Open 2 250B-A15B | 2026-07-22 |
| DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | ✓ Link | 91.6 | DeepSeek-V4-Flash Max | 2026-04-26 |
| Qwen3.8-27B | ✓ Link | 90.3 | Qwen3.8-27B | 2026-08-05 |
| Kimi K2.6 | ✓ Link | 89.6 | Kimi K2.6 | 2026-04-20 |
| Memory Transfer Learning: How Memories are Transferred Across Domains in Coding Agents | ✓ Link | 89.3 | DeepSeek V3.2 + MTL (Insight) | 2026-04-15 |
| Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning | ✓ Link | 89.0 | Nemotron 3 Ultra | 2026-06-03 |
| DeepSeek-V3.2 Technical Report | ✓ Link | 88.7 | DeepSeek-V3.2-Speciale | 2025-12-02 |
| Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity | | 87.8 | Seed2.0 Pro | 2026-06-30 |
| MAI-Thinking-1: Building a Hill-Climbing Machine | | 87.7 | MAI-Thinking-1 | 2026-06-02 |
| Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale | ✓ Link | 86.95 | Ring-2.6-1T (XHigh) | 2026-06-13 |
| Unified Audio Intelligence Without Regressing on Text Intelligence | | 85.3 | Audex 30B-A3B | 2026-07-06 |
| MiMo-V2-Flash Technical Report | ✓ Link | 85.1 | MiMo-V2-Flash (Thinking) | 2026-01-06 |
| Kimi K2.5: Visual Agentic Intelligence | ✓ Link | 85.0 | Kimi K2.5 | 2026-02-02 |
| Introducing efficiency focused "Thinking CAP" model series | ✓ Link | 84.3 | ThinkingCap-Qwen3.6-27B | 2026-07-06 |
| Qwen3.6 | ✓ Link | 83.9 | Qwen3.6-27B | 2026-04-21 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 83.6 | Qwen3.5-397B-A17B | 2026-02-16 |
| EXAONE 4.5 Technical Report | ✓ Link | 81.4 | EXAONE 4.5 (Reasoning) | 2026-04-09 |
| Mach-Mind-4-Flash Technical Report | | 80.91 | Mach-Mind-4-Flash | 2026-07-10 |
| K-EXAONE Technical Report | ✓ Link | 80.7 | K-EXAONE (236B, reasoning) | 2026-01-05 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 80.7 | Qwen3.5-27B | 2026-02-16 |
| Qwen3.6 | ✓ Link | 80.4 | Qwen3.6-35B-A3B | 2026-04-21 |
| VibeThinker-3B: Exploring the Frontier of Verifiable Reasoning in Small Language Models | ✓ Link | 80.2 | VibeThinker-3B | 2026-06-15 |
| Gemma 4 Technical Report | ✓ Link | 80.0 | Gemma 4 31B | 2026-07-02 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 78.9 | Qwen3.5-122B-A10B | 2026-02-16 |
| InCoder-32B-Thinking: Industrial Code World Model for Thinking | ✓ Link | 77.1 | InCoder-32B-Thinking | 2026-04-03 |
| Nanbeige4.1-3B: A Small General Model that Reasons, Aligns, and Acts | ✓ Link | 76.9 | Nanbeige4.1-3B | 2026-02-13 |
| Mellum2 Technical Report | ✓ Link | 75.1 | Mellum 2 Thinking | 2026-05-29 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 74.6 | Qwen3.5-35B-A3B | 2026-02-16 |
| Embarrassingly Simple Self-Distillation Improves Code Generation | ✓ Link | 74.4 | GPT-OSS-20B (high) + SSD | 2026-04-01 |
| Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale | | 74.3 | Intern-S1-Pro | 2026-03-26 |
| Solar Open Technical Report | ✓ Link | 74.2 | Solar Open (102B) | 2026-01-11 |
| Distribution-Aligned Sequence Distillation for Superior Long-CoT Reasoning | ✓ Link | 72.8 | DASD-30B-A3B-Thinking-Preview (30B MoE) | 2026-01-14 |
| Nanbeige4.2-3B: Unlocking Agentic Capabilities in a Compact Model | ✓ Link | 72.5 | Nanbeige4.2-3B | 2026-07-21 |
| IndexCache: Accelerating Sparse Attention via Cross-Layer Index Reuse | ✓ Link | 72.2 | DSA (30B) + IndexCache (1/2, training-aware) | 2026-03-12 |
| Gemma 4 Technical Report | ✓ Link | 72.0 | Gemma 4 12B Unified | 2026-07-02 |
| Introducing North Mini Code: Cohere's first model for developers | ✓ Link | 70.3 | North Mini Code | 2026-06-09 |
| Qwen3-VL Technical Report | ✓ Link | 70.1 | Qwen3-VL-235B-A22B-Thinking | 2025-11-26 |
| DiffusionGemma Technical Report | ✓ Link | 69.1 | DiffusionGemma 26B A4B | 2026-07-31 |
| Falcon-H1R: Pushing the Reasoning Frontiers with a Hybrid Model for Efficient Test-Time Scaling | ✓ Link | 68.6 | Falcon-H1R (7B) | 2026-01-05 |
| Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning | ✓ Link | 68.3 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 2025-12-23 |
| EXAONE 4.0: Unified Large Language Models Integrating Non-reasoning and Reasoning Modes | ✓ Link | 66.7 | EXAONE 4.0 32B (Reasoning) | 2025-07-15 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 65.6 | Qwen3.5-9B | 2026-02-16 |
| Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale | ✓ Link | 65.58 | Ling-2.6-1T | 2026-06-13 |
| Kwai Keye-VL-2.0 Technical Report | ✓ Link | 64.2 | Keye-VL-2.0-30B-A3B | 2026-06-09 |
| JudgeRLVR: Judge First, Generate Second for Efficient Reasoning | | 63.9 | JudgeRLVR (Qwen3-30B-A3B) | 2026-01-13 |
| Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale | ✓ Link | 62.28 | Ling-2.6-flash | 2026-06-13 |
| LFM2.5-2.6B: Deploy Agents Everywhere | ✓ Link | 59.41 | LFM2.5-2.6B | 2026-08-04 |
| Qwen3-Coder-Next Technical Report | ✓ Link | 58.93 | Qwen3-Coder-Next | 2026-02-28 |
| Introspective Diffusion Language Models | ✓ Link | 57.1 | I-DLM-32B | 2026-04-13 |
| Hunyuan Dense V1 | ✓ Link | 57 | Hunyuan-7B-Instruct | 2025-07-30 |
| Qwen3-Next | | 56.6 | Qwen3-Next-80B-A3B-Instruct | 2025-09-09 |
| How to Fine-Tune a Reasoning Model? A Teacher-Student Cooperation Framework to Synthesize Student-Consistent SFT Data | ✓ Link | 55.43 | TESSY | 2026-03-23 |
| Kimi K2: Open Agentic Intelligence | ✓ Link | 53.7 | Kimi K2 | 2025-07-28 |
| Post-Trained MoE Can Skip Half Experts via Self-Distillation | ✓ Link | 53.2 | ZEDA (Qwen3-30B-A3B) | 2026-05-18 |
| OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization | ✓ Link | 52.63 | OSCAR (GLM-4.7, 358B) | 2026-05-18 |
| Falcon-H1: A Family of Hybrid-Head Language Models Redefining Efficiency and Performance | ✓ Link | 49.71 | Falcon-H1-34B-Instruct | 2025-07-30 |
| InCoder-32B: Code Foundation Model for Industrial Scenarios | ✓ Link | 49.1 | InCoder-32B (32B) | 2026-03-17 |
| Reinforcement Learning via Self-Distillation | ✓ Link | 48.8 | SDPO (Qwen3-8B) | 2026-01-28 |
| LongCat-Flash Technical Report | ✓ Link | 48.02 | LongCat-Flash | 2025-09-01 |
| Lightning OPD: Efficient Post-Training for Large Reasoning Models with Offline On-Policy Distillation | ✓ Link | 43.9 | Qwen3-8B-Base (Lightning OPD) | 2026-04-14 |
| Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models | ✓ Link | 43.7 | Youtu-LLM 2B | 2025-12-31 |
| GlimpRouter: Efficient Collaborative Inference by Glimpsing One Token of Thoughts | ✓ Link | 43.57 | GlimpRouter (Qwen3-4B SLM + DeepSeek-R1-Distill-Qwen-32B LLM) | 2026-01-08 |
| MetaAgent-X : Breaking the Ceiling of Automatic Multi-Agent Systems via End-to-End Reinforcement Learning | | 41.0 | MetaAgent-X RL (Qwen3 8B) | 2026-05-14 |
| ERNIE 4.5 Technical Report | ✓ Link | 38.8 | ERNIE-4.5-300B-A47B | 2025-06-30 |
| Nanbeige4-3B Technical Report: Exploring the Frontier of Small Language Models | ✓ Link | 38.8 | Nanbeige4-3B-Base + OpenThoughts3 SFT (extra data) | 2025-12-06 |
| Breaking Training Bottlenecks: Effective and Stable Reinforcement Learning for Coding Models | ✓ Link | 38.7 | Qwen3-4B-Instruct (MicroCoder-GRPO) | 2026-03-08 |
| SAF-OPD: Stable Advantage Fusion for On-Policy Distillation | | 37.43 | SAF (Qwen3-8B) | 2026-07-31 |
| Trust Region On-Policy Distillation | | 36.0 | TrOPD (Qwen3-SFT-1.7B) | 2026-05-31 |
| LLaDA MoE v2: Scaling Mixture-of-Experts Diffusion Language Models | | 31.86 | LLaDA MoE v2 Base | 2026-08-04 |
| Filter, Then Reweight: Rethinking Optimization Granularity in On-Policy Distillation | ✓ Link | 28.1 | FiRe-OPD | 2026-06-01 |
| Weak-to-Strong On-Policy Distillation | ✓ Link | 27.1 | W2S-OPD (multi-teacher contrast) | 2026-07-28 |
| On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability | ✓ Link | 0.919 | Qwen3.8-Flash-Next | 2026-08-31 |
| Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe | | 0.8822 | NVIDIA Nemotron 3.5 Super VL 120B-A12B BF16 | 2026-10-15 |
| Kolibri: A Sovereign European Model on the Pareto Frontier | ✓ Link | 0.859 | Kolibri | 2026-10-03 |
| Rufus-Air: An Open LLM Post-Training Recipe | | 0.764 | Rufus-Air | 2026-09-24 |
| Granite 4.2 LLMs: How They're Built | ✓ Link | 0.7577 | Granite-4.2-30B | 2026-08-25 |
| TPPO: Tracking-Value PPO For Language Model Reasoning | ✓ Link | 0.731 | TPPO (Qwen3.5-9B) | 2026-10-12 |
| Capek 0.5: An Execution-Centric Vision-Language Model for Embodied Intelligence | | 0.7291 | Capek0.5-35B-A3B | 2026-08-07 |
| Self-Distillation with Verbal Regularization | ✓ Link | 0.721 | Qwen3-8B (VRSD single) | 2026-10-12 |
| One Model to Loop Them All: Elastic Latent Recurrence for Language Models | | 0.719 | Koi-4B | 2026-10-08 |
| MiniCPM5-2B | ✓ Link | 0.691 | MiniCPM5-2B | 2026-09-07 |
| NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness | ✓ Link | 0.6514 | NeoHorse-1-9B | 2026-09-08 |
| D^3-MOPD: Adaptive Dynamic Domain ScheDuling for Efficient Multi-Teacher Distillation | | 0.632 | D3-MOPD Qwen3.6-35B-A3B (benchmark-best checkpoint) | 2026-08-25 |
| SPADE: Self-Play in Adaptive Synthetic Executable Environments | ✓ Link | 0.494 | SPADE (Games, Qwen3-8B) | 2026-08-19 |
| ZGCM-1: A Fully Open and Extremely Efficient Foundation Model for Math and Agentic Search | ✓ Link | 0.4686 | ZGCM-1 | 2026-09-11 |
| Latent On-Policy Self-Distillation | ✓ Link | 0.458 | OLMo3-7B-Think LOPD | 2026-08-13 |
| Expert-Space Exploration in MoE Reinforcement Learning | ✓ Link | 0.42 | Qwen3-30B-A3B + ESRL | 2026-09-11 |
| MiniCPM5-1B | ✓ Link | 0.3352 | MiniCPM5-1B (Thinking) | 2026-05-25 |
| When Models Edit Too Much: On the Fidelity of Minimal Code Edits | ✓ Link | 0.332 | Minimal-edit RL (Qwen3-4B-Instruct-2507) | 2026-09-03 |
| MOPD-Router: Rethinking Teacher Routing in Multi-Teacher On-Policy Distillation | ✓ Link | 0.3086 | MOPD-Router (ExpertAlign, Qwen3-4B-Non-Thinking student, domain-labeled) | 2026-09-25 |
| Beyond Domain-Routed Supervision: Cross-Teacher Guarding for Multi-Teacher On-Policy Distillation | | 0.2957 | Road MOPD (Qwen3-4B) | 2026-10-09 |
| Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL | ✓ Link | 0.266 | Co-RL (Qwen2.5-7B, Different family+) | 2026-08-19 |
| Open-MOPD: Diagnosing and Fixing Capability Imbalance in Multi-Teacher On-Policy Distillation | ✓ Link | 0.2263 | Open-MOPD (SmolLM3-3B) | 2026-08-19 |
| Parameter Exploration for RLVR via Variational Learning | ✓ Link | 0.1517 | C3PO (OLMo-3-1025-7B) (extra data) | 2026-08-10 |