0

Meta Platforms

American tech conglomerate (Facebook, Instagram, WhatsApp); parent of FAIR, GenAI, and Meta Superintelligence Labs.

Visit
Type
frontier lab
HQ
Menlo Park, CA, USA
Founded
2004
Funding
Public (NASDAQ:META); market cap >$1.5T (2025)
Website
about.meta.com
GitHub
github.com/facebookresearch
Twitter
twitter.com/Meta

Cite

Notes

Only stored in your browser.

Evals
3
Tools
0
Models
43
Papers
0
Boards
0
People
3

Models

43

Ranked by Artificial Analysis Intelligence Index

#ModelReleasedIntelligence
1Muse Spark8 Apr 202652.2
2Llama 3.1 Instruct 405B23 Jul 202417.4
3Llama 3.3 Instruct 70B6 Dec 202414.5
4Llama 3.1 70B Instruct23 Jul 202412.5
5Llama 3.2 Instruct 90B (Vision)25 Sep 202411.9
6Llama 3.1 8B Instruct23 Jul 202411.8
7Llama 3.2 Instruct 3B25 Sep 20249.7
8Llama 2 Chat 7B18 Jul 20239.7
9Llama 3 Instruct 70B18 Apr 20248.9
10Llama 3.2 Instruct 11B (Vision)25 Sep 20248.7
11Llama 2 Chat 13B18 Jul 20238.4
12Llama 2 Chat 70B18 Jul 20238.4
13Llama 65B24 Feb 20237.4
14Llama 3 8B Instruct18 Apr 20246.4
15Llama 3.2 Instruct 1B25 Sep 20246.3
16Llama-3.2-3B18 Sep 2024-
17Llama-3.2-1B18 Sep 2024-
18Llama-3.1-8B14 Jul 2024-
19Meta-Llama-3-8B-Instruct17 Apr 2024-
20Meta-Llama-3-8B17 Apr 2024-
21Llama 2 70B--
22Llama 3 8B--
23Llama 3 405B--
24Llama 2 13B--
25Llama 1 13B--
26Llama 1 65B--
27Llama 4 Maverick 17B 128e Instruct--
28Llama 4 Scout 17B 16e Instruct--
29Llama 3 70B--
30Llama 2 34B--
31Llama 1 7B--
32Llama 3.1 405B Instruct Bf16--
33Llama 2 7B--
34CodeLlama 34B Instruct--
35CodeLlama 70B Instruct--
36Llama 13B--
37Llama 3.2 Vision 11B Instruct--
38meta-llama-3-70B--
39Llama 3.1 8B Instant--
40Llama3.2--
41Llama 3.1 405B Instruct Fp8--
42Llama 3.2 Vision 90B Instruct--
43Llama 1 33B--
43 / 43 models

Evals

3

CyberSecEval 4: Advanced Cybersecurity Evaluation Benchmarks

A suite of cybersecurity evaluation benchmarks adapted from Meta's PurpleLlama CybersecurityBenchmarks. Includes MITRE ATT&CK compliance testing, false refusal rate measurement, insecure code detection,

ActiveCybersecurity

CyberSecEval_2: Cybersecurity Risk and Vulnerability Evaluation

Assesses language models for cybersecurity risks, specifically testing their potential to misuse programming interpreters, vulnerability to malicious prompt injections, and capability to exploit known software vulnerabilities.

ActiveCybersecurity

CYBERSECEVAL 3: Advancing the Evaluation of Cybersecurity Risks and Capabilities in Large Language Models

Evaluates Large Language Models for cybersecurity risk to third parties, application developers and end users.

ActiveCybersecurity