As of December 2024, OpenAI's o1 pro is the best AI model available.

Judge as of 2024-12-25. True: in benchmark leaderboards and expert assessments from late December 2024, o1 pro (or o1 in pro mode) is the top or joint-top publicly available model across general and reasoning benchmarks (for example LiveBench, reasoning, maths and coding evaluations, Chatbot Arena). Mostly true: it leads on most reasoning-heavy benchmarks and is widely considered the strongest, but another available model beats it on some general leaderboards. Mixed: it leads on some major benchmarks and clearly trails another model (for example Gemini 2.0 or Claude 3.5 Sonnet) on others, with no consensus. Mostly false: on most major evaluations of the time another publicly available model was clearly better, and o1 pro led only in narrow areas. False: o1 pro was not available by that date, or it clearly ranked below other models on nearly all evaluations. Announced but unreleased models (for example o3) do not count as 'available'.

Set before the research began.

1. E1 Will people really pay $200 a month for OpenAI's new chatbot? | TechCrunch TechCrunch · 6 December 2024

   “OpenAI’s internal benchmarks show that o1 pro mode performs only slightly better than the standard o1 on coding and math problems:”

2. E2 o1 Turns Pro Zvi Mowshowitz (Don't Worry About the Vase, Substack) · 10 December 2024

   “Reactions to o1 Pro by professionals seem very, very positive, although it does not strictly dominate Claude Sonnet.”

3. E3 o1 Turns Pro Zvi Mowshowitz (Don't Worry About the Vase, Substack) · 10 December 2024

   “Problems o1 pro can solve that o1 can’t at all are rare; mostly it feels like things that work half the time on o1 work 90% of the time on o1 pro.”

4. E4 Introducing ChatGPT Pro - OpenAI OpenAI · 5 December 2024

   “Compared to both o1 and o1‑preview, o1 pro mode performs better on challenging ML benchmarks across math, science, and coding.”

5. E5 o1 Turns Pro Zvi Mowshowitz (thezvi.substack.com) · 10 December 2024

   “Reactions to o1 Pro by professionals seem very, very positive, although it does not strictly dominate Claude Sonnet.”

6. E6 o1 Turns Pro Zvi Mowshowitz (thezvi.substack.com) · 10 December 2024

   “Kakachia777 does a comparison of o1 Pro to Claude 3.5 Sonnet, prefers Sonnet for coding because its code is easier to maintain.”

7. E7 Simon Willison on ai and llms Simon Willison's Weblog · 7 December 2024

   “The current leaderboard has Claude 3.5 Sonnet (October edition) at the top, then various Gemini models, GPT-4o and one openly licensed model - Qwen2.5-Coder-32B - filling out the top six.”

8. E8 o1 Turns Pro Zvi Mowshowitz (thezvi.substack.com) · 10 December 2024

   “Sully: O1-Pro is probably the best model I've used for coding, hands down.”

9. E9 Best AIs of 2024 Ranked: ChatGPT vs. Gemini vs. Claude vs. Perplexity asapdrew (Substack newsletter) · 17 December 2024

   “Currently the best by a wide margin (zero competition other than o1).”

10. E10 OpenAI's New o3 & o3-mini Models asapdrew (Substack newsletter) · 21 December 2024

    “In my experience, o1-pro is currently (Dec 2024) the best AI model for advanced reasoning/logic.”

11. E11 2024 AI Model Leaderboard Year-End Rankings | Will Spurlock Will Spurlock (blog) · 27 December 2024

    “GPT-4o-latest led the December 2024 LMArena rankings with a 1366 Elo score, followed closely by Gemini-Exp-1121 at 1364.”

12. E12 AI capabilities have steadily improved over the past year Epoch AI · 30 September 2025

    “| GPQA Diamond | o1 | OpenAI | 2024-12-17 | 0.77 |”

Independent December 2024 assessments rank o1 pro as the strongest model for advanced reasoning. E10 calls it "currently (Dec 2024) the best AI model for advanced reasoning/logic", E9 says it is best "by a wide margin", and E8 quotes a user calling it the best for coding. Professional reactions were very positive, but it "does not strictly dominate Claude Sonnet", and one tester preferred Sonnet for coding (E2, E6). On general leaderboards, GPT-4o-latest and Gemini models led the December LMArena rankings (E11), and Claude 3.5 Sonnet topped another leaderboard cited in E7. That matches "mostly true": widely seen as the strongest at reasoning, but other models led some general leaderboards.