| Introducing Claude Opus 5.5 | ✓ Link | 93.9 | Claude Opus 5.5 | 2026-09-22 |
| Claude Fable 5 and Claude Mythos 5 | ✓ Link | 92.2 | Claude Mythos 5 | 2026-06-09 |
| Introducing Claude Sonnet 5.5 | ✓ Link | 90.3 | Claude Sonnet 5.5 | 2026-09-28 |
| Introducing Claude Opus 5 | ✓ Link | 89.5 | Claude Opus 5 | 2026-07-24 |
| Introducing Claude Fable 5.1 and Claude Mythos 5.1 | | 89.1 | Claude Fable 5.1 (max effort) | 2026-09-01 |
| Introducing Claude Opus 4.8 | | 84.4 | Claude Opus 4.8 | 2026-05-28 |
| Introducing Claude Haiku 5.5 | | 83.7 | Claude Haiku 5.5 (max effort) | 2026-10-07 |
| Hy4 preview | ✓ Link | 82.9 | Hy4 preview (swe-agent scaffold) | 2026-08-28 |
| Intern-S2-Preview: Scientific Agentic Foundation Model | ✓ Link | 81.67 | Intern-S2-Preview-397B (Mini-SWE-Agent) | 2026-08-13 |
| On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability | ✓ Link | 81.0 | Qwen3.8-Flash-Next (mini-SWE-agent) | 2026-08-31 |
| Introducing Claude Opus 4.7 | | 80.5 | Claude Opus 4.7 | 2026-04-16 |
| Composer 2 Technical Report | | 79.8 | Composer 2.5 (Cursor harness) | 2026-03-25 |
| Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding | ✓ Link | 78.9 | Ornith-1.0-397B (OpenHands) | 2026-06-21 |
| Introducing Laguna S 2.1 | ✓ Link | 78.5 | Laguna S 2.1 (pool) | 2026-07-21 |
| Introducing LongCat-2.0 | ✓ Link | 78.3 | LongCat-2.0 | 2026-07-05 |
| Introducing Grok 4.5 | | 78.0 | Grok 4.5 | 2026-07-08 |
| Introducing Beam: Reflection's 501B open-weight model | | 78.0 | Beam | 2026-10-05 |
| Introducing Claude Opus 4.6 | | 77.83 | Claude Opus 4.6 | 2026-02-05 |
| Kimi K2.6 | ✓ Link | 76.7 | Kimi K2.6 (In-house SWE-agent-adapted framework) | 2026-04-20 |
| The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence | ✓ Link | 76.5 | MiniMax-M2.7 | 2026-05-26 |
| Introducing Claude Opus 4.5 | | 76.2 | Claude Opus 4.5 | 2025-11-24 |
| DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | ✓ Link | 76.2 | DeepSeek-V4-Pro Max | 2026-04-26 |
| Introducing Claude Sonnet 4.6 | | 75.9 | Claude Sonnet 4.6 | 2026-02-17 |
| Hy3 | ✓ Link | 75.8 | Hy3 (SWE-agent) | 2026-07-02 |
| dots3-note Preview: A Small but Mighty Step Toward Long-Horizon Agency in Real Life | ✓ Link | 75.67 | dots3-note Preview (live-swe-agent) | 2026-08-13 |
| KAT-Coder-V2 Technical Report | | 75.4 | KAT-Coder-V2 (Claude Code) | 2026-03-29 |
| Composer 2 Technical Report | | 73.7 | Composer 2 (Cursor harness) | 2026-03-25 |
| GLM-5: From Vibe Coding to Agentic Engineering | ✓ Link | 73.3 | GLM-5 (OpenHands) | 2026-02-11 |
| DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence | ✓ Link | 73.3 | DeepSeek-V4-Flash Max | 2026-04-26 |
| Kimi K2.5: Visual Agentic Intelligence | ✓ Link | 73.0 | Kimi K2.5 (Internal agent scaffold) | 2026-02-02 |
| Ling-3.0-flash | | 72.4 | Ling-3.0-flash (OpenHands) | 2026-07-23 |
| MiMo-V2-Flash Technical Report | ✓ Link | 71.7 | MiMo-V2-Flash (Thinking) | 2026-01-06 |
| Ornith-1.5: From Self-Scaffolding to Self-Improvement | | 71.4 | Ornith-1.5-35B-A3B (OpenHands) | 2026-08-20 |
| Qwen3.6 | ✓ Link | 71.3 | Qwen3.6-27B (Internal agent scaffold) | 2026-04-21 |
| DeepSeek-V3.2 Technical Report | ✓ Link | 70.2 | DeepSeek-V3.2 (Internal framework) | 2025-12-02 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 69.3 | Qwen3.5-397B-A17B (Internal agent scaffold) | 2026-02-16 |
| Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding | ✓ Link | 69.3 | Ornith-1.0-35B (OpenHands) | 2026-06-21 |
| MindForge: Teaching Small Language Models Whole-Life-Cycle Software Engineering via Source-Free Program Synthesis | | 67.77 | MindForge-27B (Mini-SWE-Agent) | 2026-07-29 |
| Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe | | 67.72 | NVIDIA Nemotron 3.5 Super VL 120B-A12B BF16 | 2026-10-15 |
| Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning | ✓ Link | 67.7 | Nemotron 3 Ultra (OpenHands) | 2026-06-03 |
| Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters | ✓ Link | 67.4 | Step-3.5-Flash | 2026-02-11 |
| Qwen3.6 | ✓ Link | 67.2 | Qwen3.6-35B-A3B (Internal agent scaffold) | 2026-04-21 |
| Qwen3-Coder-Next Technical Report | ✓ Link | 64.3 | Qwen3-Coder-Next (OpenHands) | 2026-02-28 |
| Laguna M.1/XS.2 Technical Report | ✓ Link | 63.1 | Laguna M.1 (pool) | 2026-05-26 |
| Introducing Laguna XS 2.1 | ✓ Link | 63.1 | Laguna XS 2.1 (pool) | 2026-07-02 |
| Self-Evolving Coding Agents via Harness-Policy Optimization | ✓ Link | 59.7 | Self-HPO-Qwen3.5-35B-A3B | 2026-10-16 |
| One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agents | ✓ Link | 59.0 | Logics-SWE-Qwen3.6-27B (SWE-agent) | 2026-09-20 |
| Laguna M.1/XS.2 Technical Report | ✓ Link | 57.7 | Laguna XS.2 (pool) | 2026-05-26 |
| Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding | ✓ Link | 52.0 | Ornith-1.0-9B (OpenHands) | 2026-06-21 |
| Dockerless: Environment-Free Program Verifier for Coding Agents | | 50.0 | Dockerless-RL-9B (OpenHands) | 2026-06-26 |
| Kimi K2: Open Agentic Intelligence | ✓ Link | 47.3 | Kimi K2 | 2025-07-28 |
| NVIDIA Nemotron 3: Efficient and Open Intelligence | ✓ Link | 45.78 | Nemotron 3 Super (OpenHands) | 2025-12-24 |
| Granite 4.2 LLMs: How They're Built | ✓ Link | 41.89 | Granite-4.2-30B | 2026-08-25 |
| Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem | ✓ Link | 40.0 | ROME (30B MoE, activated 3B) (iFlow CLI) | 2025-12-31 |
| NVIDIA Nemotron 3.5 Lightning | ✓ Link | 39.33 | NVIDIA Nemotron 3.5 Lightning 30B-A3B (NeMo Evaluator) | 2026-08-11 |
| Scaling Embeddings Outperforms Scaling Experts in Language Models | | 38.1 | LongCat-Flash-Lite (68.5B) | 2026-01-29 |
| LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling | ✓ Link | 31.0 | LoopCoder-v2 | 2026-06-16 |
| SWE-smith: Scaling Data for Software Engineering Agents | ✓ Link | 8.4 | SWE-agent-LM-32B | 2025-04-30 |