egov.mn

Anthropic vs. OpenAI: Who’s Winning the AI Race?

Anthropic and OpenAI have traded the title of top AI model eight times since 2024. See who leads now, and how long they've held it.

This story was prepared automatically by software.

Anthropic vs. OpenAI: Who’s Winning the AI Race?

Who’s Ahead in the Race Between Anthropic and OpenAI?

Key Takeaways

  • Anthropic’s Fable 5.2 leads the Artificial Analysis Intelligence Index with a score of 64.6 as of Nov. 17, 2026, ahead of OpenAI’s GPT-6.2 Astra at 62.1.

  • Anthropic has held the lead for seven consecutive months since April 2026, its longest stretch in front, after leadership changed hands eight times between September 2024 and April 2026.

  • The two companies released 13 frontier models in 2026 through mid-November, more than the 11 released in 2024 and 2025 combined.

The rivalry between Anthropic and OpenAI has become one of the defining competitions in artificial intelligence. Each new flagship model offers another chance to claim the top spot on AI intelligence benchmarks.

This graphic tracks the highest scores achieved by each company on the Artificial Analysis (AA) Intelligence Index over time. It includes Anthropic’s Opus and Fable models and OpenAI’s leading GPT and o-series models, showing only releases that set a new high for their company.

The index currently combines nine tests covering knowledge work, coding, science, and general knowledge. AA updated its methodology on Oct. 27, so scores before and after that date are not directly comparable.

Anthropic Has Had the Smartest AI Models Since April 2026

For much of the AI race, neither Anthropic nor OpenAI could hold onto the top spot for long. Between May 2025 and April 2026 alone, leadership changed seven times, with some advantages lasting only weeks.

That pattern changed when Anthropic’s Opus 4.7 overtook OpenAI in April 2026. Despite a steady stream of new releases from both companies, Anthropic has remained ahead on the Artificial Analysis Intelligence Index ever since.

The table below lists every model on the chart in order of release, along with its score on each version of the index:

Release DateCompanyAI ModelIntelligence Index Score by Version

v4.3.2v5.0.1

Mar 14, 2023OpenAIGPT-46.70

Mar 4, 2024AnthropicOpus 38.72

May 13, 2024OpenAIGPT-4o7.33

Sep 12, 2024OpenAIo1-preview11.38

Dec 5, 2024OpenAIo115.23

Apr 16, 2025OpenAIo320.20

May 22, 2025AnthropicOpus 420.64

Aug 5, 2025AnthropicOpus 4.122.85

Aug 7, 2025OpenAIGPT-522.98

Nov 13, 2025OpenAIGPT-5.124.74

Nov 24, 2025AnthropicOpus 4.529.10

Dec 11, 2025OpenAIGPT-5.230.45

Feb 5, 2026AnthropicOpus 4.631.95

Mar 5, 2026OpenAIGPT-5.438.98

Apr 16, 2026AnthropicOpus 4.740.69

May 28, 2026AnthropicOpus 4.841.79

Jun 9, 2026AnthropicFable 549.63

Jul 9, 2026OpenAIGPT-5.6 Sol46.97

Jul 24, 2026AnthropicOpus 550.78

Sep 1, 2026AnthropicFable 5.153.35

Sep 3, 2026OpenAIGPT-6 Astra52.6756.85

Sep 22, 2026AnthropicOpus 5.557.6259.11

Nov 9, 2026AnthropicFable 5.264.56

Nov 17, 2026OpenAIGPT-6.2 Astra62.14

Scores for models released before Opus 4.8 (May 2026) are Artificial Analysis estimates. Version 5.0.1 scores are shown for models plotted after the Oct. 27 methodology change.

Some victories in the AI race have been remarkably short-lived. Anthropic’s Opus 4.5 held the lead for just 17 days before OpenAI’s GPT-5.2 overtook it in December 2025.

The longer-term picture is more balanced than the current streak suggests. Since March 2024, Anthropic has spent 529 days in front, compared with 459 days for OpenAI. That makes its current run especially significant in a competition where leadership has historically been difficult to maintain.

Anthropic’s lead also survived AA’s Oct. 27 update, which added two tests, upgraded three, and removed three.

However, different benchmarks can produce different winners. On Epoch AI’s Capabilities Index, OpenAI’s GPT-6.2 Astra leads with a score of 174.2, narrowly ahead of Fable 5.2 at 173.5. The difference falls within both models’ margins of error.

Frontier AI Model Releases Have Doubled in Pace This Year

The competition has also accelerated. In 2025, Anthropic and OpenAI released a new frontier model roughly every 52 days. In 2026, that interval has shortened to about 25 days, with 13 releases through mid-November.

Anthropic has driven much of that pace, accounting for eight of the 13 releases this year. OpenAI has often responded within days. GPT-6 Astra arrived two days after Fable 5.1 and scored 0.7 points lower. In November, GPT-6.2 Astra followed Fable 5.2 by eight days and finished 2.4 points behind.

But benchmark leadership is only one measure of success. OpenAI also faces growing competition in the broader AI chatbot market, where user adoption tells a different story.

ChatGPT accounted for 88% of global AI chatbot website traffic at the end of 2024, but its share fell to 56% by September 2026 as Anthropic’s Claude and Google Gemini gained ground.

Learn More on the Voronoi App

If you enjoyed today’s post, check out Ranked: The Smartest AI Models in 2026 on Voronoi.

Share
Next storyI let Apple’s always-listening AI summarize my life for two weeks
Newsletter

Get the latest news delivered to your inbox.

Related articles