CNMO 2024
Ring-1T 88.5% leads since October 2025 
DeepSeek-V3 43.2% leads since December 2024 Ring-1T 88.5% leads since October 2025 Models answer problems from the 2024 Chinese National High School Mathematics Olympiad. DeepSeek reports Pass@1 averaged over 16 samples.
DeepSeek built the eval set from the contest and did not release its prepared version or problem count.
- Released
- December 2024
- Kind
- Competition problems with answers
- Level
- Competition
- Scored
- Checked automatically
- Answers
- Short answer
- Strongest model tested
- Kimi K2 Jul 2025 #116 on the capability index
- Last in a launch post
- July 2025, dropped after it saturated
- Last model added
- January 2026
- Scores refreshed
- November 12th, 2026
Ring-1T leads with 88.5% among models released by January 2026
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
DeepSeek-V3 leads with 43.2% across all independent runs
Best score per model across the official leaderboard, Epoch AI’s own runs and any board Epoch AI copied. Each bar and row names its source.
Ring-1T leads with 88.5%, claims included
Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model.
The newest model with a score here came out in January 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 16 more
- 1 Ring-1T 88.5%
- 2 DeepSeek-R1 May 2025 86.9%
- 3
Step 3 83.7% - 4 DeepSeek-R1 78.8%
- 5 STEP3-VL-10B 78.2%
- 6 Kimi K2 Jul 2025 74.3%
- 7 DeepSeek-V3 43.2%
Source: Labs’ model cards
- 1 DeepSeek-V3 43.2%
- 2
Qwen2.5-72B 15.9% - 3 Claude 3.5 Sonnet (October 2024) 13.1%
- 4 GPT-4o May 2024 10.8%
- 5 DeepSeek-V2.5 Sep 2024 10.8%
- 6
Llama 3.1-405B 6.80% - 7 DeepSeek-V2 (MoE-236B, May 2024) 2.80%
Source: Official CNMO 2024 leaderboard
- 1 Ring-1T 88.5%
- 2 DeepSeek-R1 May 2025 86.9%
- 3 Step 3 83.7%
- 4 DeepSeek-R1 78.8%
- 5 STEP3-VL-10B 78.2%
- 6 Kimi K2 Jul 2025 74.3%
- 7 DeepSeek-V3 43.2%
- 8 Qwen2.5-72B 15.9%
- 9 Claude 3.5 Sonnet (October 2024) 13.1%
- 10 GPT-4o May 2024 10.8%
- 11 DeepSeek-V2.5 Sep 2024 10.8%
- 12 Llama 3.1-405B 6.80%
- 13 DeepSeek-V2 (MoE-236B, May 2024) 2.80%
Sources: Labs’ model cards and official CNMO 2024 leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Ring-1T | 88.5% | Oct 2025 | Lab model card | |
| 2 | DeepSeek-R1 May 2025 | 86.9% | May 2025 | Lab model card | |
| 3 | Step 3 | 83.7% | Jul 2025 | Lab model card | |
| 4 | DeepSeek-R1 | 78.8% | Jan 2025 | Lab model card | |
| 5 | STEP3-VL-10B | 78.2% | Jan 2026 | Lab model card | |
| 6 | Kimi K2 Jul 2025 | 74.3% | Jul 2025 | Lab model card | |
| 7 | DeepSeek-V3 | 43.2% | Dec 2024 | Lab model card |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | DeepSeek-V3 | 43.2% | Dec 2024 | Official leaderboard | |
| 2 | Qwen2.5-72B | 15.9% | Sep 2024 | Official leaderboard | |
| 3 | Claude 3.5 Sonnet (October 2024) | 13.1% | Oct 2024 | Official leaderboard | |
| 4 | GPT-4o May 2024 | 10.8% | May 2024 | Official leaderboard | |
| 5 | DeepSeek-V2.5 Sep 2024 | 10.8% | Sep 2024 | Official leaderboard | |
| 6 | Llama 3.1-405B | 6.80% | Jul 2024 | Official leaderboard | |
| 7 | DeepSeek-V2 (MoE-236B, May 2024) | 2.80% | May 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Ring-1T | 88.5% | Oct 2025 | Lab model card | |
| 2 | DeepSeek-R1 May 2025 | 86.9% | May 2025 | Lab model card | |
| 3 | Step 3 | 83.7% | Jul 2025 | Lab model card | |
| 4 | DeepSeek-R1 | 78.8% | Jan 2025 | Lab model card | |
| 5 | STEP3-VL-10B | 78.2% | Jan 2026 | Lab model card | |
| 6 | Kimi K2 Jul 2025 | 74.3% | Jul 2025 | Lab model card | |
| 7 | DeepSeek-V3 | 43.2% | Dec 2024 | Official leaderboard | |
| 8 | Qwen2.5-72B | 15.9% | Sep 2024 | Official leaderboard | |
| 9 | Claude 3.5 Sonnet (October 2024) | 13.1% | Oct 2024 | Official leaderboard | |
| 10 | GPT-4o May 2024 | 10.8% | May 2024 | Official leaderboard | |
| 11 | DeepSeek-V2.5 Sep 2024 | 10.8% | Sep 2024 | Official leaderboard | |
| 12 | Llama 3.1-405B | 6.80% | Jul 2024 | Official leaderboard | |
| 13 | DeepSeek-V2 (MoE-236B, May 2024) | 2.80% | May 2024 | Official leaderboard |
The benchmark's own leaderboard has 7 models. Switch the scores above to rank them together.
4 models have held number one in labs’ own numbers since its release in December 2024 2 models have held number one across all independent runs since its release in December 2024 5 models have held number one with claims included since its release in December 2024
Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and what labs claim in their cards.
Sources: Official CNMO 2024 leaderboard and labs’ model cards
DeepSeek’s own number for DeepSeek-R1 passed 85% in May 2025
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Sources: Official CNMO 2024 leaderboard and labs’ model cards
Labs dropped it from launch posts once the best score passed 85%
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| Jul 31, 2025 | Step 3 | 83.7 | CNMO24 |
| Jul 11, 2025 | Kimi K2 | 74.3 | CNMO 2024, Avg@16 |
| May 28, 2025 | DeepSeek-R1 | 86.9 | CNMO 2024 (Pass@1) |
| Mar 16, 2025 |  ERNIE 4.5 | CNMO2024 | |
| Jan 20, 2025 | DeepSeek-R1 | 78.8 | CNMO 2024 (Pass@1) |
| Dec 26, 2024 | DeepSeek-V3 | 43.2 | CNMO 2024 (Pass@1) |
5 system cards report it, from 4 labs
Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Jan 14, 2026 | STEP3-VL-10B technical report | 78.2 | CNMO2024, SeRe (Tables 2 and 3). PaCoRe 81.17 | Capabilities |
| Oct 21, 2025 | Ring-1T technical report | 88.5 | CNMO 2024 (Table 1) | Capabilities |
| Jul 21, 2025 | Kimi K2 technical report | 74.3 | CNMO 2024 (Avg@16) (Table 3) | Capabilities |
| Jan 20, 2025 | DeepSeek-R1 technical report | 78.8 | CNMO 2024 (Pass@1), DeepSeek-R1 (Table 8, same as final column of Table 3) | Capabilities |
| Dec 26, 2024 | DeepSeek-V3 technical report | 43.2 | CNMO 2024 (Pass@1), DeepSeek-V3 (Table 6) | Capabilities |
One lab claim next to the leaderboard
Each score a lab reported for its own model, beside the same model's best score on the leaderboard. Gaps mostly come from different harnesses, effort levels and task subsets. 1 of them match, which usually means one side copied the other. Every benchmark.
| Model | Lab says | Board says | Gap |
|---|---|---|---|
| DeepSeek-V3 | 43.2% CNMO 2024 (Pass@1) | 43.2% Leaderboard | Same |
More in Math
- AIME 2026 In use Models solve the 30 integer-answer problems from the 2026 AIME I and II math competitions. Scored by exact answer match. GPT-5.5 100%
- IMO-AnswerBench In use 400 olympiad problems with short verifiable answers, robustified against memorization. Scored by answer match. GLM-5.2 91%
- HMMT 2026 In use Models solve the 33 final-answer problems from the February 2026 Harvard-MIT Mathematics Tournament. Scored by exact answer match. GPT-5.5 98.5%
- PolyMath In use Models solve 9,000 parallel math problems in 18 languages across four difficulty levels, from school math to frontier math. Final answers are checked. Qwen3.7-Max 86.5%
- MathArena (Apex, IMO, USAMO) In use Evaluates models on newly released math competitions after release. Final-answer or human/LLM-judged proof scores. Claude Opus 4.8 81.3%
- MathArena Apex Shortlist In use Models solve 47 hard final-answer problems that MathArena selected from 2025 competitions, where the best models scored about 50% at selection. Scored by accuracy over repeated attempts. GPT-5.5 98.4%