Debugging

Models 20

• Gen 7
  Laguna S 2.1

  By poolside

  Laguna S 2.1 is a 118B parameter Mixture of Experts model with 8B activated parameters per token, built for agentic coding and long horizon software engineering work. It has a 1,048,576 token context window, native interleaved reasoning, and mixed global/sliding window attention across 48 layers.

  💻Coding✉️Codebase Q&A🐞Debugging🔬Code testing🤖Agents🔧Code refactoring
  NewText

  Released 1d ago
• Gen 2 SWE
  SWE 1.7

  By Cognition AI

  SWE 1.7 is Cognition's most capable coding model to date, reaching frontier level intelligence at a much lower cost. Trained with reinforcement learning from a Kimi K2.7 base, it is optimized for long horizon, asynchronous agentic software engineering tasks and thorough codebase exploration.

  💻Coding✉️Codebase Q&A🐞Debugging🔧Code refactoring🔬Test generation💻Vibe coding
  NewCoding

  Released 14d ago
• Gen 2 Mistral
  Leanstral 1.5 119B A6B

  By Mistral AI

  Leanstral 1.5 119B A6B is a mixture of experts model with 119 billion total parameters and 6.5 billion active per token, built by Mistral AI as a code agent for formal verification in Lean 4. It accepts text and image input and produces text output, with a 256k token context window. It solves competition math benchmarks like miniF2F and PutnamBench and can verify code correctness properties, uncovering real bugs in open source repositories.

  📐Mathematical proofs🐞Debugging🤖Agents🔢Math
  NewCoding

  Released 20d ago
• Gen 7 Ornith
  Ornith 1.0 397B FP8

  By DeepReinforce

  Ornith-1.0-397B-FP8 is DeepReinforce's open-source 397B MoE agentic coding model with FP8 quantization. Built on Qwen 3.5 and RL-trained to jointly optimize scaffolds and solutions, it achieves state-of-the-art performance on Terminal-Bench 2.1, SWE-Bench, and NL2Repo. Supports reasoning, tool use, and 262K context. MIT licensed.

  💻Coding🐞Debugging🤖Agents
  NewText

  Released 22d ago
• Gen 3 Ornith
  Ornith 1.0 397B

  By DeepReinforce

  Ornith-1.0-397B is a 397B MoE open-source reasoning model for agentic coding, post-trained on Qwen 3.5 MoE via a self-improving RL framework that jointly learns task solutions and the scaffolds guiding them. Achieves 82.4 on SWE-Bench Verified and 77.5 on Terminal-Bench 2.1. MIT licensed with tool-calling and 256K context support.

  🔍Code analysis🐞Debugging📚Large Language Models🤖Agents💻Vibe coding
  NewMultimodal

  Released 25d ago
• Closed betaGen 7 GPT
  GPT 5.6 Terra

  By OpenAI

  GPT 5.6 Terra is a balanced model in OpenAI's GPT 5.6 preview series. It offers performance competitive with GPT 5.5 while costing about half as much. Terra shows improved coding and cybersecurity capability that scales with reasoning effort, and is paired with a layered safety and misuse safeguard stack during its limited preview release.

  💬Chatting🐞Debugging💻Coding🔒Cybersecurity🔒AI security
  NewText

  Released 26d ago
• Closed betaGen 7 GPT
  GPT 5.6 Luna

  By OpenAI

  GPT 5.6 Luna is the fast and affordable tier of OpenAI's GPT 5.6 preview series. It is the lowest cost model in the family while still delivering strong capability, including cybersecurity gains that scale with reasoning effort. Luna is released as part of a limited, safeguard paired preview alongside the Sol and Terra models.

  💬Chatting🐞Debugging💻Coding🔒Cybersecurity🔒AI security
  NewText

  Released 26d ago
• Gen 2
  North Mini Code 1.0

  By Cohere

  North Mini Code 1.0 is Cohere’s Apache 2.0 agentic coding model with 30B total parameters and 3B active parameters, built for repo-level software engineering, terminal agents, local coding, and code generation.

  💻Coding🐞Debugging🤖Agents🔧Code refactoring💻Vibe coding
  NewCoding

  Released 1mo ago
• Gen 2 MAI
  MAI Code 1 Flash

  By Microsoft

  MAI-Code-1-Flash is Microsoft AI’s 5B inference-efficient agentic coding model integrated into GitHub Copilot, VS Code, and the Microsoft stack.

  💻Coding🐞Debugging🔧Code optimization🔧Code refactoring💻Vibe coding
  NewCoding

  Released 1mo ago
• Gen 3 Grok
  Grok Build 0.1

  By xAI

  Grok Build 0.1 is xAI’s fast coding model for agentic coding, web development, debugging, MCP support, and tool-calling workflows.

  💻Coding🐞Debugging🤖Agents🎯Code autocompletion💻Vibe coding
  NewMultimodal

  Released 1mo ago
• Gen 3 Composer
  Composer 2.5

  By Anysphere

  Composer 2.5 is Cursor’s coding-agent model for long-running software tasks, complex instruction following, and agentic coding inside Cursor.

  💻Coding🐞Debugging🔧Code refactoring🔬Test generation💻Vibe coding
  NewMultimodal

  Released 2mo ago
• Gen 3
  Laguna XS 2

  By poolside

  Laguna XS.2 is poolside’s open 33B MoE coding model for agentic software work on local machines. It uses 3B activated parameters per token, supports native reasoning with interleaved thinking around tool calls, and is designed for long-horizon coding with low KV-cache requirements and Mac-friendly deployment.

  💻Coding🐞Debugging🤖Agents🤔Logical reasoning💻Software engineer assistance💻Vibe coding
  NewMultimodal

  Released 2mo ago
• Gen 7
  Carnice V2 27B

  By Kai Stephens

  Carnice-V2-27B is a BF16 supervised fine-tune of Qwen3.6-27B for Hermes-style agent traces. It is built for agentic conversational use, instruction following, and tool-oriented workflows, and is released as a fully merged standalone checkpoint rather than only a LoRA adapter.

  🤖Agents🐞Debugging💻Coding🔍Image interpretation
  NewText

  Released 2mo ago
• Gen 7 SWE
  SWE 1.6

  By Cognition AI

  SWE-1.6 is Cognition’s latest model for software engineering agents, built to improve both raw capability and what the company calls “model UX.” It is positioned as fast, low-cost, and smoother to use in practice, with up to 950 tok/s and better behavior in agent workflows.

  💻Coding🔍Code reviews🐞Debugging
  Text

  Released 3mo ago
• Gen 2 GPT
  GPT 5.1 Codex

  By OpenAI

  GPT 5.1 Codex is a code-focused variant that maps specs to working code, writes diffs and tests, and understands repositories with long context, tool calling, and strict JSON or patch output.

  💻Coding🐞Debugging💘Dating advice🐍Python lessons💻Vibe coding
  Coding

  Released 8mo ago
• Gen 7 Qwen
  Qwen3 Coder

  By Alibaba

  Qwen3 Coder is Alibaba’s code-focused Qwen3 variant. It handles repo-aware generation, completion, debugging, and test creation, with long context, tool and function calling, and strict JSON or diff outputs for IDEs and agents.

  💻Coding🔎Problem solving🐞Debugging🔍Document search
  Text

  Released 1y ago
• Gen 2 V0
  v0 1.5 lg

  By Vercel

  v0-1.5-lg is Vercel's larger composite AI coding model for full-stack web app generation. It pairs a bigger frontier base LLM with retrieval-augmented context and a custom AutoFix error-correction model, supports a 512K token context window, and excels at complex reasoning tasks like specialized physics engines and multi-step database migrations.

  💻Coding🐞Debugging💻Web development assistance🔨Ui code generation💻Vibe coding
  Coding

  Released 1y ago
• Gen 2 V0
  v0 1.5 md

  By Vercel

  v0-1.5-md is a composite AI coding model from Vercel that generates full-stack web application code. It pairs a frontier base LLM with retrieval-augmented context and a custom AutoFix model that corrects errors during code streaming, producing higher-quality React and Next.js code with fewer mistakes.

  🐞Debugging💻Web development assistance🔨Ui code generation💻Vibe coding
  Coding

  Released 1y ago
• Gen 2 V0
  v0 1.0 md

  By Vercel

  v0-1.0-md is Vercel's original composite AI coding model for the v0 platform. It combines Anthropic's Claude Sonnet 3.7 as its base model with retrieval-augmented context and a custom AutoFix error-correction pipeline to generate full-stack web application code from text and image prompts.

  💻Coding🐞Debugging💻Web development assistance🔨Ui code generation💻Vibe coding
  Coding

  Released 1y ago
• Gen 7 DeepSeek
  DeepSeek-R1

  By DeepSeek

  DeepSeek R1 is a reasoning-first large language model built to solve complex problems with explicit multi-step thinking. It excels at math, coding, and logical analysis, supports long context, tool/function calling, and structured JSON outputs, and can trade latency for higher accuracy via extended "thinking" budgets.

  💻Coding📊Data analysis🐞Debugging🇯🇵Japanese lessons
  Text

  Released 1y ago