ModelTape
All modelsNew releasesRankingsRuns on my GPUOpen the app →
ModelTape / New releases
Last 30 days

New LLM releases

38 new language models in the last 30 days, 5 of them this week. 27 have open weights you can download and run yourself.

ModelLabReleasedSizeECIWeights
Nex N2.6Nex-agi2026-11-10321B MoE–open
Claude Fable 5.2Anthropic2026-11-09–173.4closed
Muse EmberMeta2026-11-09––closed
KLM 1.6MinimaLabs2026-11-090.1B–open
Cola2ByteDance2026-11-068.8B–open
mirageCactus-Compute2026-11-05––open
S5 nanoMaritaca-ai2026-11-024.7B–open
Raon 1 nanoKRAFTON2026-10-293.5B–open
Gemini 4 CarbonGoogle2026-10-28–169.8closed
Particle-1Particle2026-10-28748B–closed
SenseNova Loop 9B PreviewSenseNova2026-10-289.7B–open
MAI-Cyber-1.1-FlashMicrosoft2026-10-27––closed
Ling 3.1 flash 1019inclusionAI2026-10-27572B MoE–open
GPT-6.1 TerraOpenAI2026-10-26–167.7closed
LBS 27BinclusionAI2026-10-2627B–open
Kimi K3.1 ProMoonshot AI2026-10-232.8T161.4open
BobaHumans&2026-10-22––closed
WeVisOne VL 4BTencent2026-10-224.8B–open
WeVisOne VL 2BTencent2026-10-222.4B–open
Mirendil-BaseMirendil2026-10-21––closed
ERNIE 5.5 Flash BaseBaidu2026-10-21286B MoE–open
flintRecursiveSI2026-10-2140B MoE–open
ERNIE 5.5 Flash Base BF16Baidu2026-10-21286B MoE–open
AmpersandHumans&2026-10-20744B–closed
MiniMax-M3.1-FlashMiniMax2026-10-19220B MoE–open
MatterSeek New Materials Reverse DesignYuyaoTechnology2026-10-19––open
Ling 3.0 flash SanteinclusionAI2026-10-19127B MoE–open
DeepSeek V4.1 Flash 1017DeepSeek2026-10-17552B MoE159.4open
DeepSeek V4 Flash Vision ExpDeepSeek2026-10-17305B MoE–open
Mercury 3Inception2026-10-16––closed
Axion 1.0 35BAxion1-02026-10-1635B MoE–open
NVIDIA Nemotron Labs Dory1.0 4B BF16NVIDIA2026-10-164B–open
NVIDIA Nemotron Labs Dory1.0 4B MXFP8NVIDIA2026-10-164B–open
Muse Spark OpenMeta2026-10-15822B MoE155.4open
MAI-Base-2Microsoft2026-10-152.8T–closed
Hy Memory 30B A3BTencent2026-10-1530B MoE–open
granite switch 4.1 8bIBM2026-10-148.8B–open
granite switch 4.1 30bIBM2026-10-1430B–open
Data updated Thu, 12 Nov 2026 09:42 UTC. Benchmarks, Capabilities Index and model announcements: Epoch AI, “Capabilities & benchmarking” and “Data on AI models” (CC BY 4.0); category scores are ModelTape's own fit on Epoch's results. Releases and prices: models.dev (MIT). Sizes and quantizations: Hugging Face. “Runs on” figures are estimates from model size, quantization and memory bandwidth (±30–50%).
Live app · All models · New releases · Rankings · 6 GB · 8 GB · 12 GB · 16 GB · 24 GB · 32 GB · Legal notice (Impressum) · Privacy