Timeline

2024-01-08
Mixtral — Mistral AI sparse-MoE full-attention text-only open-weights large frontier-generalist
2024-03-28
Jamba — AI21 Labs hybrid hybrid text-only open-weights large frontier-generalist
2024-04-22
Phi-3 — Microsoft dense full-attention text-only undisclosed medium on-device
2024-07-23
Llama 3 — Llama Team, AI @ Meta dense full-attention text-only open-weights frontier frontier-generalist
2024-07-31
Llama 3.1 — Meta (Llama Team, AI @ Meta) dense full-attention text-only open-weights frontier frontier-generalist
2024-12-12
Phi-4 — Microsoft Research dense full-attention text-only undisclosed medium reasoning-specialised
2024-12-27
DeepSeek-V3 — DeepSeek-AI sparse-MoE full-attention text-only open-weights frontier frontier-generalist
2024-12-31
OLMo 2 — OLMo Team, Allen Institute for AI (Ai2) dense full-attention text-only open-weights-open-data medium frontier-generalist
2025-01-22
DeepSeek-R1 — DeepSeek-AI sparse-MoE full-attention text-only open-weights frontier reasoning-specialised
2025-03-17
xLSTM-7B — NXAI (NX-AI) dense SSM text-only open-weights medium frontier-generalist
2025-03-25
Gemma 3 — Google DeepMind (Gemma Team) dense efficient-attention multimodal open-weights large on-device
2025-05-15
Qwen3 — Qwen Team dense full-attention text-only open-weights large frontier-generalist
2025-05-15
Qwen3 — Qwen Team sparse-MoE full-attention text-only open-weights frontier frontier-generalist
2025-07-28
Kimi K2 — Kimi Team (Moonshot AI) sparse-MoE full-attention text-only open-weights frontier frontier-generalist
2025-08-05
gpt-oss — OpenAI sparse-MoE hybrid text-only open-weights frontier reasoning-specialised
2025-08-08
GLM-4.5 (ARC series) — Zhipu AI & Tsinghua University (GLM-4.5 Team) sparse-MoE full-attention text-only open-weights frontier frontier-generalist
2025-08-08
GLM-4.5 (ARC series) — Zhipu AI & Tsinghua University (GLM-4.5 Team) sparse-MoE full-attention text-only open-weights frontier frontier-generalist
2025-11-01
Kimi Linear — Kimi Team (Moonshot AI) sparse-MoE hybrid text-only open-weights large frontier-generalist
2025-12-02
DeepSeek-V3.2 — DeepSeek-AI sparse-MoE efficient-attention text-only open-weights reasoning-specialised
2025-12-15
OLMo 3 — Allen Institute for AI (Olmo Team) dense hybrid text-only open-weights-open-data large frontier-generalist
2025-12-23
Nemotron 3 Nano — NVIDIA sparse-MoE SSM text-only open-weights-open-data large reasoning-specialised
2026-01-06
MiMo-V2 — Xiaomi (LLM-Core) sparse-MoE efficient-attention text-only open-weights frontier reasoning-specialised
2026-01-29
LongCat-Flash — Meituan LongCat Team sparse-MoE text-only open-weights large long-context
2026-02-13
Nanbeige4.1 — Nanbeige LLM Lab, Boss Zhipin dense text-only open-weights medium frontier-generalist
2026-02-17
GLM-5 — Zhipu AI & Tsinghua University (GLM-5 Team) sparse-MoE efficient-attention text-only open-weights frontier frontier-generalist
2026-02-19
Trinity — Arcee AI sparse-MoE hybrid text-only open-weights frontier frontier-generalist
2026-02-28
Qwen3-Coder-Next — Qwen Team hybrid hybrid text-only open-weights large domain-specific
2026-03-12
Tiny Aya — Cohere Labs / Cohere dense hybrid text-only open-weights medium on-device
2026-04-03
Nemotron 3 Super — NVIDIA sparse-MoE hybrid text-only open-weights-open-data frontier reasoning-specialised
2026-04-21
Qwen3.5 — Qwen Team hybrid hybrid text-only undisclosed frontier frontier-generalist
2026-04-26
DeepSeek-V4 — DeepSeek-AI sparse-MoE efficient-attention text-only open-weights frontier long-context
2026-05-06
ZAYA1 — Zyphra sparse-MoE efficient-attention text-only undisclosed medium reasoning-specialised
2026-05-26
MiniMax-M2 series — MiniMax sparse-MoE full-attention multimodal open-weights frontier frontier-generalist
2026-05-29
Mellum 2 — JetBrains (with Constructor University, Bremen) sparse-MoE hybrid text-only open-weights large domain-specific
2026-06-09
Nemotron 3 Ultra (Nemotron 3 family) — NVIDIA sparse-MoE hybrid text-only open-weights-open-data frontier frontier-generalist
2026-06-11
MiniMax M3 — MiniMax (with authors from Peking University, NVIDIA, Zhejiang University, HUST, Nanjing University, Hangzhou Dianzi University) sparse-MoE efficient-attention multimodal open-weights frontier long-context
2026-06-13
Ling 2.6 — Ling Team, Inclusion AI sparse-MoE efficient-attention text-only open-weights frontier frontier-generalist
2026-06-15
VibeThinker — Weibo dense full-attention text-only undisclosed medium reasoning-specialised
2026-07-22
Solar Open 2 — Upstage (Upstage Solar Team) sparse-MoE hybrid text-only open-weights frontier frontier-generalist
2026-07-24
Gemma 4 — Google DeepMind (Gemma Team) dense efficient-attention multimodal open-weights large on-device
2026-07-27
Kimi K3 — Kimi Team (Moonshot AI) sparse-MoE hybrid multimodal open-weights frontier frontier-generalist