Advanced reasoning

Models 14

• Gen 3 Qwen
  Qwen 3.6 Plus Preview

  By Alibaba

  Qwen3.6 Plus Preview is a preview large language model in the Qwen Plus line, built with a hybrid architecture aimed at better efficiency, scalability, reasoning, and agentic behavior. It is positioned as especially strong for coding, front-end development, and complex problem-solving, and is offered on OpenRouter with a 1,000,000-token context window.

  💬Chatting🔍Advanced reasoning🧩Complex problem solving🖥️Frontend development💻Conversational coding
  NewMultimodal

  Released 2mo ago
• Gen 3
  Alpamayo 1.5 10B

  By NVIDIA

  Alpamayo 1.5-10B is NVIDIA’s open 10B vision-language-action model for autonomous driving. It is built as a steerable reasoning engine for AV research, combining multi-camera visual input, text, and egomotion history to produce both chain-of-causation reasoning and future driving trajectories.

  🚗Autonomous driving🔍Advanced reasoning🔍Image interpretation👁️Computer vision assistance
  Multimodal

  Released 3mo ago
• Gen 7
  Nemotron Cascade 2 30B A3B

  By NVIDIA

  Nemotron-Cascade 2 is an open 30B Mixture-of-Experts model (3B activated) trained with Cascade RL and multi-domain on-policy distillation for strong reasoning and agentic capabilitie

  💬Chatting🔍Advanced reasoning💻Conversational coding⌨️Competitive programming coaching
  Text

  Released 3mo ago
• Gen 3 Mistral
  Mistral Small 4

  By Mistral AI

  Mistral Small 4 is an open hybrid model that unifies instruct, reasoning, and coding in a single multimodal model with a 256k context window.

  💬Chatting🔍Advanced reasoning🔍Image interpretation🔢Math💻Conversational coding
  Multimodal

  Released 3mo ago
• Gen 7
  Nemotron 3 Super

  By NVIDIA

  Nemotron 3 Super is NVIDIA’s open Mixture-of-Experts hybrid Mamba-Transformer model designed for high-throughput agentic workloads and long-context reasoning.

  🤖Agents📚Large Language Models🔍Advanced reasoning💻Conversational coding
  Text

  Released 3mo ago
• Gen 3
  Covenant 72B

  By Covenant AI

  Covenant-72B is a permissionless collaboratively trained 72B-parameter language model trained from scratch via decentralized coordination on Bittensor infrastructure.

  📚Large Language Models💬Chatting🔍Advanced reasoning
  Text

  Released 3mo ago
• Gen 3 GPT
  GPT 5.4 Pro

  By OpenAI

  GPT-5.4 Pro is the maximum-performance variant of GPT-5.4 for people who want the strongest results on complex tasks in ChatGPT and the API.

  💬Chatting🤖Agents🔍Advanced reasoning🧩Complex problem solving🔍Web research summarization💻Conversational coding
  Multimodal

  Released 3mo ago
• Gen 3 GPT
  GPT 5.4 Thinking

  By OpenAI

  GPT-5.4 Thinking is the ChatGPT-facing GPT-5.4 model mode tuned for professional work, emphasizing stronger reasoning, coding, and agentic workflows with fewer back-and-forth turns.

  💬Chatting🤖Agents🔍Advanced reasoning🧮Advanced math tutoring🧩Complex problem solving💻Conversational coding
  Multimodal

  Released 3mo ago
• Gen 7
  Nanbeige4.1 3B

  By Nanbeige LLM Lab

  Nanbeige4.1-3B is a 3B-parameter reasoning model optimized to combine strong multi-step reasoning, preference alignment, and agentic tool use in a compact size.

  🔍Advanced reasoning🔍Information retrieval🧩Complex problem solving🎓Adaptive tutoring
  Text

  Released 4mo ago
• Gen 7
  Nemotron 3 Nano 30b

  By NVIDIA

  Nemotron-3-Nano-30B-A3B is NVIDIA’s 30B-class LLM trained from scratch as a unified model for both reasoning and standard chat tasks. It can be configured to produce a reasoning trace before the final answer, letting developers switch reasoning behavior through the chat template.

  💬Chatting💻Coding🔍Advanced reasoning💻Conversational coding
  Text

  Released 5mo ago
• Gen 3
  Kimi K2 Thinking

  By Moonshot AI

  Kimi K2 Thinking is Moonshot AI’s open-source reasoning-oriented K2 variant, built for deep multi-step thinking and tool use. It uses native INT4 quantization, supports a 256k context window, and is aimed at long-horizon agentic workflows rather than fast reflex chat.

  💬Chatting💻Coding🔍Advanced reasoning🧩Complex problem solving
  Text

  Released 7mo ago
• Gen 3
  Llama 3.3 Nemotron Super 49B

  By NVIDIA

  Llama-3.3-Nemotron-Super-49B-v1.5 is NVIDIA’s upgraded 49B reasoning model derived from Meta Llama-3.3-70B-Instruct. It is post-trained for reasoning, chat preference alignment, and agentic tasks such as RAG and tool use.

  💬Chatting💻Coding🔍Advanced reasoning💻Conversational coding
  Text

  Released 8mo ago
• Gen 3
  Seed OSS 36B Instruct

  By ByteDance

  Seed-OSS-36B-Instruct is ByteDance Seed’s open-source 36B instruction model built for long-context work, reasoning, agents, and general-purpose developer use. ByteDance positions the Seed-OSS line as efficient for its scale, reporting strong open-benchmark results from 12T training tokens.

  💬Chatting💻Coding🔍Advanced reasoning🛠️Comprehensive assistance
  Multimodal

  Released 10mo ago
• Gen 7
  Kimi K2 Instruct

  By Moonshot AI

  Kimi-K2-Instruct-0905 is Moonshot AI’s upgraded Kimi K2 instruct model, a 1T-parameter MoE system with 32B active parameters. It is tuned for agentic coding, better real-world coding performance, and stronger frontend generation without the long-thinking behavior of the dedicated thinking variant

  💬Chatting🔍Code reviews🔍Advanced reasoning🖥️Frontend development💻Conversational coding
  Text

  Released 11mo ago