0

Anthropic

Frontier AI safety lab building the Claude family of models, founded by former OpenAI researchers.

Visit
Type
frontier lab
HQ
San Francisco, CA, USA
Founded
2021
Funding
Private, ~$300B+ valuation 2025; backed by Google, Amazon, Lightspeed, Menlo
Website
anthropic.com
GitHub
github.com/anthropics
Twitter
twitter.com/AnthropicAI

Cite

Notes

Only stored in your browser.

Evals
3
Tools
1
Models
42
Papers
1
Boards
0
People
47

Models

42

Ranked by Artificial Analysis Intelligence Index

#ModelReleasedIntelligence
1Claude Opus 5.522 Sep 202657.6
2Claude Sonnet 5.528 Sep 202656
3Claude Fable 5.11 Sep 202653.4
4Claude Opus 524 Jul 202650.8
5Claude Fable 59 Jun 202649.6
6Claude Haiku 5.57 Oct 202643.4
7Claude Opus 4.828 May 202641.8
8Claude Opus 4.716 Apr 202640.7
9Claude Sonnet 530 Jun 202638.2
10Claude Opus 4.65 Feb 202631.9
11Claude Sonnet 4.617 Feb 202630.1
12Claude Opus 4.524 Nov 202523.7
13Claude Sonnet 4.529 Sep 202519.3
14Claude 4.1 Opus5 Aug 202518.6
15Claude 4 Sonnet22 May 202516.6
16Claude 4 Opus22 May 202516.6
17Claude 4.5 Haiku15 Oct 202515.4
18Claude Sonnet 3.724 Feb 202515.3
19Claude 3.5 Haiku22 Oct 20248.9
20Claude Opus 34 Mar 20248.7
21Claude 3.5 Sonnet22 Oct 20247.9
22Claude 3.5 Sonnet (June '24)21 Jun 20247.2
23Claude Sonnet 34 Mar 20245.9
24Claude 3 Haiku4 Mar 20245.6
25Claude 2.121 Nov 20235.6
26Claude 2.011 Jul 20235.5
27Claude Instant14 Mar 20235
28Claude Sonnet 4 5 Search--
29Claude Opus 4 1 Search--
30Claude Haiku 4.5--
31Claude Opus 4 6 Search--
32Claude Sonnet 4 6 Search--
33Claude Sonnet 4 20250514 Thinking--
34Claude Opus 4 1 20250805 Thinking--
35Claude Sonnet 5 Search--
36Claude Opus 4 5 Search--
37Claude Opus 4 Search--
38Claude Opus 4.1--
39Claude Opus 4 20250514 Thinking--
40Claude Opus 4--
41Claude Sonnet 4--
42Claude Mythos Preview--
42 / 42 models

Evals

3

Sycophancy Eval

Evaluate sycophancy of language models across a variety of free-form text-generation tasks.

ActiveAssistants

Anthropic HH-RLHF

Anthropic's helpful & harmless preference dataset - paired human-rated assistant responses widely used both as a preference-training corpus and as a reward-model benchmark.

ActiveSafetyHarmful ContentInstruction Following

Agentic Misalignment: How LLMs could be insider threats

Eliciting unethical behaviour (most famously blackmail) in response to a fictional company-assistant scenario where the model is faced with replacement.

ActiveScheming

Tools

1

Anthropic HH-RLHF

Anthropic's foundational helpful-and-harmless human preference dataset - the first major public RLHF corpus and a long-time community baseline.

PreferenceSafetyJailbreak ResistanceMulti Turn Dialog

Papers

1

Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

preprint

2022