AI models built for step-by-step thinking: maths, science, planning, agents and hard coding problems. Compare reasoning quality, speed, thinking-token cost and whether open weights are available.
NNNVIDIA Nemotron 3 Nano Omni 30B A3B Reasoning is an open-weight 31B multimodal MoE that understands text, images,…
Cohere Command A Reasoning is a 111B enterprise reasoning model with 256K context for agents, tool use, RAG…
Q8Qwen3.8-2.4T-A95B is Qwen’s open-weight flagship MoE with 2.4T total and 95B active parameters, a 1M-token context and hosted…
GOgpt-oss-20b is OpenAI’s open-weight 21B (3.6B active) reasoning model for fast local and private deployments, with a 131K…
DVDeepSeek-V4-Flash-Vision-Exp is DeepSeek’s experimental vision-enabled V4 Flash for multimodal agents and visual coding, with a 1M context and…
Claude Mythos 5.1 is Anthropic’s invitation-only model for long-horizon research in areas like cybersecurity and biology, with a…
NNNVIDIA Nemotron 3.5 Lightning 30B A3B is an open-weight 30B (3B active) MoE for fast sub-agents and long-running…
NNNVIDIA Nemotron 3 Super 120B A12B is an open-weight 120B (12B active) reasoning model for agents, RAG and…
MSMuse Spark 1.3 is Meta’s model for long-horizon agentic coding, computer use and multi-agent workflows, available in public…
Claude Opus 4.6 is Anthropic’s February 2026 Opus model for complex agents and long-horizon work, with a 1M…
Claude Opus 4.7 is Anthropic’s model for complex reasoning and agentic coding, with a 1M context, 128K output…
Claude Fable 5 is Anthropic’s model for demanding reasoning and long-horizon agents, with always-on adaptive thinking, a 1M…
GOgpt-oss-120b is OpenAI’s open-weight 117B (5.1B active) reasoning model that fits on a single H100 GPU, with a…
G5GPT-5.5 Pro is the higher-compute GPT-5.5 tier for the hardest reasoning and precision-critical professional work, priced at $30/M…
NNNVIDIA Nemotron 3 Ultra 550B A55B is NVIDIA’s 550B (55B active) frontier reasoning model for agents, coding, planning…
DeepSeek-R1 is the January 2025 MIT-licensed reasoning model with 671B total, 37B active parameters and 128K context. Its…