MT-AIME2024
Qwen3-235B-A22B 80.8% leads since April 2025 Qwen3-235B-A22B 80.8% leads since April 2025 Models solve the 30 AIME 2024 problems translated into 55 languages. A rule-based verifier checks each final answer.
It is a subset of the MCLM multilingual math benchmark.
- Made by
- Yonsei University, OneLineAI, and KAIST AI
- Released
- February 2025
- Kind
- Reasoning in many languages
- Level
- Competition
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Strongest model tested
- o3-mini #124 on the capability index
- Last in a launch post
- April 2025, dropped before it saturated
- Last model added
- May 2025
- Scores refreshed
- November 12th, 2026
Qwen3-235B-A22B leads with 80.8% among models released by April 2025
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.
o3-mini leads with 45.3% 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.
Qwen3-235B-A22B leads with 80.8%, 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 April 2025, 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 Qwen3-235B-A22B 80.8%
- 2 Qwen3-32B 75%
- 3 Qwen3-30B-A3B 73.9%
- 4 Qwen3-14B 73.3%
- 5 Qwen3-8B 65.4%
- 6 Qwen3-4B 60.7%
- 7 Qwen3-1.7B 36.1%
Source: Labs’ model cards
Source: Official MT-AIME2024 leaderboard
- 1 Qwen3-235B-A22B 80.8%
- 2 Qwen3-32B 75%
- 3 Qwen3-30B-A3B 73.9%
- 4 Qwen3-14B 73.3%
- 5 Qwen3-8B 65.4%
- 6 Qwen3-4B 60.7%
- 7 o3-mini 55-language aggregate 45.3%
- 8 Qwen3-1.7B 36.1%
- 9
DeepSeek-R1-Distill-Qwen-32B 55-language aggregate 31% - 10 DeepSeek-R1-Distill-Qwen-7B 55-language aggregate 26.6%
- 11 DeepSeek-R1-Distill-Qwen-1.5B 55-language aggregate 17.2%
Sources: Labs’ model cards and official MT-AIME2024 leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen3-235B-A22B | 80.8% | Apr 2025 | Lab model card | |
| 2 | Qwen3-32B | 75% | May 2025 | Lab model card | |
| 3 | Qwen3-30B-A3B | 73.9% | May 2025 | Lab model card | |
| 4 | Qwen3-14B | 73.3% | May 2025 | Lab model card | |
| 5 | Qwen3-8B | 65.4% | May 2025 | Lab model card | |
| 6 | Qwen3-4B | 60.7% | May 2025 | Lab model card | |
| 7 | Qwen3-1.7B | 36.1% | May 2025 | Lab model card |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | o3-mini | 45.3% | 55-language aggregate | Jan 2025 | Official leaderboard |
| 2 | DeepSeek-R1-Distill-Qwen-32B | 31% | 55-language aggregate | Jan 2025 | Official leaderboard |
| 3 | DeepSeek-R1-Distill-Qwen-7B | 26.6% | 55-language aggregate | Jan 2025 | Official leaderboard |
| 4 | DeepSeek-R1-Distill-Qwen-1.5B | 17.2% | 55-language aggregate | Jan 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Qwen3-235B-A22B | 80.8% | Apr 2025 | Lab model card | |
| 2 | Qwen3-32B | 75% | May 2025 | Lab model card | |
| 3 | Qwen3-30B-A3B | 73.9% | May 2025 | Lab model card | |
| 4 | Qwen3-14B | 73.3% | May 2025 | Lab model card | |
| 5 | Qwen3-8B | 65.4% | May 2025 | Lab model card | |
| 6 | Qwen3-4B | 60.7% | May 2025 | Lab model card | |
| 7 | o3-mini | 45.3% | 55-language aggregate | Jan 2025 | Official leaderboard |
| 8 | Qwen3-1.7B | 36.1% | May 2025 | Lab model card | |
| 9 | DeepSeek-R1-Distill-Qwen-32B | 31% | 55-language aggregate | Jan 2025 | Official leaderboard |
| 10 | DeepSeek-R1-Distill-Qwen-7B | 26.6% | 55-language aggregate | Jan 2025 | Official leaderboard |
| 11 | DeepSeek-R1-Distill-Qwen-1.5B | 17.2% | 55-language aggregate | Jan 2025 | Official leaderboard |
The benchmark's own leaderboard has 4 models. Switch the scores above to rank them together.
Number one over time Number one over time 2 models have held number one with claims included since its release in February 2025
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 MT-AIME2024 leaderboard and labs’ model cards
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Sources: Official MT-AIME2024 leaderboard and labs’ model cards
No launch post has listed it since April 2025
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 |
|---|---|---|---|
| Apr 29, 2025 | Qwen3-235B-A22B | 80.8 | MT-AIME2024 |
7 system cards report it, from 1 lab
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 |
|---|---|---|---|---|
| May 14, 2025 | Qwen3-8B technical report | 65.4 | MT-AIME2024, thinking mode (Table 17). Non-thinking 16.6 (Table 18) | Capabilities |
| May 14, 2025 | Qwen3-4B technical report | 60.7 | MT-AIME2024, thinking mode (Table 17). Non-thinking 13.9 (Table 18) | Capabilities |
| May 14, 2025 | Qwen3-32B technical report | 75 | MT-AIME2024, thinking mode (Table 13). Non-thinking 24.1 (Table 14) | Capabilities |
| May 14, 2025 | Qwen3-30B-A3B technical report | 73.9 | MT-AIME2024, thinking mode (Table 15). Non-thinking 24.6 (Table 16) | Capabilities |
| May 14, 2025 | Qwen3-235B-A22B technical report | 80.8 | MT-AIME2024, thinking mode (Table 11). Non-thinking 32.4 (Table 12) | Capabilities |
| May 14, 2025 | Qwen3-14B technical report | 73.3 | MT-AIME2024, thinking mode (Table 15). Non-thinking 23.2 (Table 16) | Capabilities |
| May 14, 2025 | Qwen3-1.7B technical report | 36.1 | MT-AIME2024, thinking mode (Table 19). Non-thinking 4.9 (Table 20) | Capabilities |
More in Multilingual
- MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
- WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET.
K-EXAONE 90.5% - MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language. Qwen3.7-Max 87%
- CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories.
Solar Open 2 90.7% - Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy.
Nemotron 3 Ultra 550B-A55B Base 90.1% - MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%