Large Language Models
Mistral AI
Mistral (Large 4)
The French AI lab's model lineup — frontier flagships alongside some of the most-used open models in the industry.
Anthropic
Claude (Opus 5.5)
Anthropic's model family, built with a heavy emphasis on Constitutional AI — currently the Fable 5.1, Opus 5.5, Sonnet 5.5, and Haiku 5.5 lineup.
xAI
Grok (4.7)
xAI's model family, notable for tight integration with X (formerly Twitter) and one of the largest disclosed training clusters in the industry.
OpenAI
GPT (6)
OpenAI's flagship model family — the engine behind ChatGPT, now the GPT-6 generation (Astra, Sol, Luna and Instant).
Google DeepMind
Gemini (3.8 Flash / 3.1 Pro)
Google DeepMind's natively multimodal model family, notable for very long context windows and strong reasoning benchmarks.
Meta Superintelligence Labs
Muse Spark (1.3)
Meta's first proprietary, closed-weight flagship model — a striking departure from the open-weight Llama line.
Z.ai (Zhipu AI)
GLM (5.3)
Z.ai's open-weight GLM family — general-purpose reasoning and coding models from China.
Tencent
Hunyuan (Hy4 Preview)
Tencent's Hunyuan family — large open-source mixture-of-experts models for coding, productivity and science.
IBM
Granite (4.2)
IBM's open Granite models — compact, Apache 2.0 models for business, coding and agent workloads.
Alibaba
Qwen (3.8)
Alibaba's model family, shipped in an unusually wide range of sizes from mobile-friendly to a 2.4T-parameter flagship.
DeepSeek
DeepSeek (V4)
The Chinese open-weight model family that shook up the industry by matching closed frontier models at a fraction of the reported training cost.
Upstage
Solar (Pro 4)
Upstage's Solar models — Korean-built enterprise and document-focused language models.
Moonshot AI
Kimi (K3)
Moonshot AI's open-weight Kimi family — long-context reasoning and coding models from China.
Preferred Networks
PLaMo (3.0 Prime)
Preferred Networks' PLaMo — a Japanese language model developed from scratch in Japan, available by API or on-premises.
NVIDIA
Nemotron (3 Ultra)
NVIDIA's open Nemotron models — large mixture-of-experts models built for long-running AI agents and enterprise deployment.
MiniMax
MiniMax (M3)
MiniMax's flagship model family — open-weight, multimodal and aimed at agents and coding.
Cohere
Command (A+)
Cohere's Command family — enterprise models built for retrieval-augmented generation, citations and secure deployment.
Baidu
ERNIE (5.1)
Baidu's ERNIE family — Chinese-language and multimodal models, with older generations released as open weights.
Rakuten
RakutenAI (3.0)
Rakuten's Japanese-optimised flagship — a ~700B-parameter open-weight model from the GENIAC project.
Microsoft
Phi (4)
Microsoft's small language models — compact, MIT-licensed models designed to run on consumer hardware.
Sarvam AI
Sarvam (105B)
Sarvam AI's open models trained from scratch in India for the country's 22 official languages.
ByteDance
Seed (Doubao-Seed 2.0)
ByteDance's Seed models behind the Doubao assistant — Pro, Lite, Mini and Code variants offered through Volcano Engine.
ELYZA
ELYZA (Diffusion)
ELYZA's Japanese models — including a Japanese-specialised diffusion language model.
Technology Innovation Institute (TII)
Falcon (H1R)
TII's open Falcon models from Abu Dhabi — efficient hybrid Mamba-Transformer models, including Arabic-focused versions.
AI21 Labs
Jamba (2)
AI21 Labs' Jamba models — hybrid SSM-Transformer models built for long context, now shifting to AI21's Maestro platform.
Amazon
Nova (2 Pro)
Amazon's Nova models on AWS — multimodal enterprise models, now being refocused on a new frontier-model effort.
Naver
HyperCLOVA X (SEED)
Naver's HyperCLOVA X — Korean-language models, with open SEED releases including a native omnimodal model.
Allen Institute for AI (AI2)
OLMo (3.1)
AI2's fully open language models — weights, data, code and training recipes are all public.
AI Singapore
SEA-LION (v4)
AI Singapore's SEA-LION — open models tuned for Southeast Asian languages and contexts.
Meta
Llama (4 Scout / Maverick)
Meta's open-weight model family — widely used as the base for countless fine-tuned and self-hosted AI products.