| Gemini 3.1 Pro: A smarter model for your most complex tasks | | 59 | Gemini 3.1 Pro | 2026-02-19 |
| Kimi K3: Open Frontier Intelligence | ✓ Link | 58.7 | Kimi K3 (Max) | 2026-07-27 |
| Introducing Muse Spark Open, The Most Capable US Open-Weight Model to Date | ✓ Link | 58.5 | Muse Spark Open | 2026-10-15 |
| Kimi K2.6 | ✓ Link | 52.2 | Kimi K2.6 | 2026-04-20 |
| KAT-Coder-V2.5 Technical Report | | 50.3 | KAT-Coder-V2.5 | 2026-07-06 |
| Introducing Beam: Reflection's 501B open-weight model | | 49.7 | Beam | 2026-10-05 |
| Intern-S2-Preview: Scientific Agentic Foundation Model | ✓ Link | 49.11 | Intern-S2-Preview-397B | 2026-08-13 |
| Kimi K2.5: Visual Agentic Intelligence | ✓ Link | 48.7 | Kimi K2.5 | 2026-02-02 |
| Inkling-Small | ✓ Link | 48.7 | Inkling-Small | 2026-07-27 |
| Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe | | 47.13 | NVIDIA Nemotron 3.5 Super VL 120B-A12B BF16 | 2026-10-15 |
| The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence | ✓ Link | 47.0 | MiniMax-M2.7 | 2026-05-26 |
| Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning | ✓ Link | 44.6 | Nemotron 3 Ultra | 2026-06-03 |
| Scaling the Horizon, Not the Parameters: Reaching Trillion-Parameter Performance with a 35B Agent | ✓ Link | 44.3 | Agents-A1 | 2026-06-29 |
| Sage-A1.5: From Long-Horizon Execution to Scientific Exploration with Capability Composition | ✓ Link | 44.3 | Sage-A1.5 | 2026-10-20 |
| Muse Glimmer 30B: An On-Device Agentic Model | ✓ Link | 43.6 | Muse Glimmer-30B (High) | 2026-10-17 |
| Magenta: Closing the Loop Between Mathematical Reasoning and Lean Verification | | 42.7 | K2-Horizon-375B-A23B | 2026-09-10 |
| NVIDIA Nemotron 3: Efficient and Open Intelligence | ✓ Link | 42.05 | Nemotron 3 Super | 2025-12-24 |
| GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models | ✓ Link | 41.7 | GLM-4.5 (355B) | 2025-08-08 |
| Motif 3: Technical Report | ✓ Link | 40.6 | Motif 3 | 2026-08-10 |
| Granite 4.2 LLMs: How They're Built | ✓ Link | 38.76 | Granite-4.2-30B | 2026-08-25 |
| Introducing North Mini Code: Cohere's first model for developers | ✓ Link | 38.2 | North Mini Code | 2026-06-09 |
| GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models | ✓ Link | 37.3 | GLM-4.5-Air | 2025-08-08 |
| Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation | | 36.4 | Nemotron-Cascade-2-30B-A3B | 2026-03-19 |
| Nanbeige4.2-3B: Unlocking Agentic Capabilities in a Compact Model | ✓ Link | 35.6 | Nanbeige4.2-3B | 2026-07-21 |
| Unified Audio Intelligence Without Regressing on Text Intelligence | | 35.2 | Audex 30B-A3B | 2026-07-06 |
| Closing the Loop: Practical Training Recipes for Looped Language Models | | 35.11 | Dory1.0-4B | 2026-09-30 |
| TPPO: Tracking-Value PPO For Language Model Reasoning | ✓ Link | 35.0 | TPPO (Qwen3.5-9B) | 2026-10-12 |
| Quantization-Aware Healing: A Practical Recipe for Recovering Compressed, 4-Bit LLMs | | 34.2 | Hypernova-60B | 2026-08-21 |
| One Model to Loop Them All: Elastic Latent Recurrence for Language Models | | 33.4 | Koi-4B | 2026-10-08 |
| Nemotron 3 Nano: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning | ✓ Link | 33.3 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 2025-12-23 |
| LFM2.5-24B-A2B: Built for Fast On-Device Agents | | 32.71 | LFM2.5-24B-A2B | 2026-10-13 |
| NVIDIA Nemotron 3.5 Lightning | ✓ Link | 32.6 | NVIDIA Nemotron 3.5 Lightning 30B-A3B | 2026-08-11 |
| Nemotron 3 Nano Omni: Efficient and Open Multimodal Intelligence | ✓ Link | 32.0 | Nemotron 3 Nano Omni | 2026-04-27 |
| Introducing K2 Horizon: Frontier Performance, Radically Open | ✓ Link | 31.6 | K2-Horizon-7B | 2026-09-03 |
| AgentMercury: Your Agent Can Synthesize Verifiable Environments for Business Scenarios at scale | | 29.9 | Qwen3.5-35B-A3B + GRPO + Ours | 2026-08-21 |