How AI actually works, and the models behind it

The foundation models behind today's AI tools, and the concepts behind how they work — tracked, benchmarked, and explained in plain language.

0
models tracked
0
labs represented
0
explainers
Jul 2026
last updated

Models

9 of 9 models
xAI logo

xAI

CLOSED WEIGHT

Grok (4.5)

xAI's model family, notable for tight integration with X (formerly Twitter) and one of the largest disclosed training clusters in the industry.

Released Jul 2026128K tokens
53
Artificial Analysis Intelligence Index
See more →
Anthropic logo

Anthropic

CLOSED WEIGHT

Claude (Sonnet 5)

Anthropic's model family, built with a heavy emphasis on Constitutional AI — currently the Sonnet 5, Opus 4.8, Fable 5, and Haiku 4.5 lineup.

Released Jun 20261M tokens
69.2%
SWE-bench Pro (Opus 4.8)
61.4
Artificial Analysis Intelligence Index (Opus 4.8)
See more →
DeepSeek logo

DeepSeek

OPEN WEIGHT

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.

Released Apr 20261M tokens
90.8%
MMLU (V3/R1 generation)
See more →
OpenAI logo

OpenAI

CLOSED WEIGHT

GPT (GPT-5.5)

OpenAI's flagship model family — the engine behind ChatGPT's default experience.

Released Apr 20261M tokens (API)
92.4%
MMLU
88.7%
SWE-bench Verified
See more →
Meta Superintelligence Labs logo

Meta Superintelligence Labs

CLOSED WEIGHT

Muse Spark

Meta's first proprietary, closed-weight flagship model — a striking departure from the open-weight Llama line.

Released Apr 2026128K tokens
Mistral AI logo

Mistral AI

CLOSED WEIGHT

Mistral (Medium 3.5)

The French AI lab's model lineup — closed-weight flagships alongside some of the most-used open models in the industry.

Released Apr 2026128K tokens
92.0%
HumanEval
See more →
Google DeepMind logo

Google DeepMind

CLOSED WEIGHT

Gemini (3.1 Pro)

Google DeepMind's natively multimodal model family, notable for very long context windows and strong reasoning benchmarks.

Released Mar 20261M tokens
94.3%
GPQA Diamond
77.1%
ARC-AGI-2
See more →
Alibaba logo

Alibaba

OPEN WEIGHT

Qwen (3.5)

Alibaba's open-weight model family, shipped in an unusually wide range of sizes from mobile-friendly to frontier-scale.

Released Feb 2026128K tokens
86.1%
MMLU (2.5 generation)
See more →
Meta logo

Meta

OPEN WEIGHT

Llama (4 Scout / Maverick)

Meta's open-weight model family — widely used as the base for countless fine-tuned and self-hosted AI products.

Released Apr 2025128K tokens
86.0%
MMLU
88.4%
HumanEval
See more →