LLM Registry

Makers 24

Anthropic28›OpenAI34›Google37›Meta16›DeepSeek31›Alibaba (Qwen)31›Moonshot AI (Kimi)15›Z.ai (Zhipu / GLM)17›xAI (Grok)15›Mistral AI34›Cohere17›MiniMax10›ByteDance3›Tencent2›Thinking Machines Lab2›Poolside1›Ornith3›Ant Group (inclusionAI)2›NVIDIA1›IBM3›Apodex AI2›Multiverse Computing1›Institute of Foundation Models4›HUMAIN1›

Anthropic

Claude Fable / Mythos (Mythos-class)5›Claude Opus8›Claude Sonnet8›Claude Haiku3›Claude (legacy 1–2.x)4›

Claude Fable / Mythos (Mythos-class)Anthropic · 5 versions

Claude Fable 5.1current2026-09-01Mythos-class model released 1 September 2026; 1M context, 128K max output, $10/$50 per MTok with cache reads cut 75% to $0.25. Scores 55.8% on Terminal-Bench 4.0.ProprietaryProprietaryReasoningFrontier1Mspecs, benchmarks & pricing →Claude Mythos 5.12026-09-01Same model as Claude Fable 5.1 with lighter safeguards, released 1 September 2026; invitation-only through Project Glasswing for cybersecurity and life-sciences work.ProprietaryProprietaryReasoningFrontier1Mspecs, benchmarks & pricing →Claude Fable 52026-06-09First widely-released Mythos-class model; 1M context, 128K max output, $10/$50 per MTok. Always-on thinking; ships with safety classifiers that fall…ProprietaryProprietaryReasoningFrontierspecs, benchmarks & pricing →Claude Mythos 52026-06-09Same underlying model as Fable 5 with some safeguards lifted; available only through Project Glasswing (cyberdefense/infrastructure partners, with th…ProprietaryProprietaryFrontierspecs, benchmarks & pricing →Claude Mythos Preview2026-04Invitation-only early Mythos-class model used by a small number of organizations for cybersecurity work; superseded by Mythos 5.ProprietaryProprietaryFrontierOn-devicespecs, benchmarks & pricing →

A registry of large language models — frontier and open-weight. Pick a maker, then a line, to browse every version; each version has its own page with specs, benchmarks, pricing, APIs and lineage. Numbers trace to verified sources and refresh as models ship.