ICML 2025 Experts Benchmark
Models answer tasks submitted by participants at ICML 2025 in Vancouver. Kaggle reports overall performance and individual task outcomes.
- Made by
- Kaggle
- Released
- July 2025
- Kind
- Language reasoning
- Level
- University
- Scored
- Settled by events
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Written by crowdworkers
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- o4-mini #95 on the capability index
- Last model added
- June 2025
- Scores refreshed
- November 20th, 2026
Gemini 2.5 Pro leads with 81.1% among models released by June 2025
Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.
The newest model with a score here came out in June 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1
Gemini 2.5 Pro Jun 2025 81.1%
- 2 o4-mini 75.7%
- 3 Claude Sonnet 4 70.3%
- 4
Gemini 2.0 Flash Feb 2025 64.9%
- 5 Claude 3.5 Sonnet 56.8%
- 6
DeepSeek-R1 May 2025 51.4%
Source: Official ICML 2025 Experts Benchmark leaderboard
| # | Model | Score | Setting | Released | Run on | Source |
|---|---|---|---|---|---|---|
| 1 | | 81.1% | Jun 2025 | Jul 2025 | Official leaderboard | |
| 2 | o4-mini | 75.7% | Apr 2025 | Jul 2025 | Official leaderboard | |
| 3 | Claude Sonnet 4 | 70.3% | May 2025 | Jul 2025 | Official leaderboard | |
| 4 | | 64.9% | Feb 2025 | Jul 2025 | Official leaderboard | |
| 5 | Claude 3.5 Sonnet | 56.8% | Jun 2024 | Jul 2025 | Official leaderboard | |
| 6 | | 51.4% | May 2025 | Jul 2025 | Official leaderboard |
Rankings by part
We read 37 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Generate ML conference networking questions, Zeno's Paradox: Flashing Light and Reversed Fibonacci sequence. Claude 3.5 Sonnet leads 24 of them, and 4 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Generate ML conference networking questions
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Zeno's Paradox: Flashing Light
6 models we track have a score here.
- 1 o4-mini ↑1 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
DeepSeek-R1 May 2025 ↑3 100%
- 4
Gemini 2.5 Pro Jun 2025 ↓3 100%
- 5 Claude 3.5 Sonnet 0%
- 6
Gemini 2.0 Flash Feb 2025 ↓2 0%
Reversed Fibonacci sequence
6 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑3 100%
- 2 o4-mini 100%
- 3
DeepSeek-R1 May 2025 ↑3 100%
- 4
Gemini 2.5 Pro Jun 2025 ↓3 100%
- 5 Claude 3.5 Sonnet 0%
- 6 Claude Sonnet 4 ↓3 0%
Water Jug Riddle
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Refuse Harmful Request
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Logic Puzzle: 1 = 5
6 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑3 100%
- 2 Claude 3.5 Sonnet ↑3 0%
- 3 o4-mini ↓1 0%
- 4 Claude Sonnet 4 ↓1 0%
- 5
DeepSeek-R1 May 2025 ↑1 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Atbash Cipher Encoding
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Generate a novel stable crystal structure
6 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑3 100%
- 2 o4-mini 100%
- 3
Gemini 2.5 Pro Jun 2025 ↓2 100%
- 4 Claude 3.5 Sonnet ↑1 0%
- 5 Claude Sonnet 4 ↓2 0%
- 6
DeepSeek-R1 May 2025 0%
ARC-AGI Puzzle: 00576224
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4
DeepSeek-R1 May 2025 ↑2 100%
- 5
Gemini 2.5 Pro Jun 2025 ↓4 100%
- 6 Claude Sonnet 4 ↓3 0%
Checking Current FEV1 Reference Equations
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 0%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 0%
- 3 o4-mini ↓1 0%
- 4 Claude Sonnet 4 ↓1 0%
- 5
DeepSeek-R1 May 2025 ↑1 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Short sci-fi story with female lead
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
Gemini 2.5 Pro Jun 2025 ↓4 100%
- 6
DeepSeek-R1 May 2025 0%
Find an analytical solution for 2^x + 3x = 11
6 models we track have a score here.
- 1 o4-mini ↑1 100%
- 2
Gemini 2.5 Pro Jun 2025 ↓1 100%
- 3 Claude 3.5 Sonnet ↑2 0%
- 4
Gemini 2.0 Flash Feb 2025 0%
- 5 Claude Sonnet 4 ↓2 0%
- 6
DeepSeek-R1 May 2025 0%
Identify best quadrotor control algorithm
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
10th number in pi
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Word Chain Generator
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2 o4-mini 100%
- 3
Gemini 2.5 Pro Jun 2025 ↓2 100%
- 4
Gemini 2.0 Flash Feb 2025 0%
- 5 Claude Sonnet 4 ↓2 0%
- 6
DeepSeek-R1 May 2025 0%
Give an example of a palindrome
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Parse VCF file into a pandas DataFrame
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Riddle: Magic Square Age
6 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 100%
- 2 Claude 3.5 Sonnet ↑3 0%
- 3
Gemini 2.0 Flash Feb 2025 ↑1 0%
- 4 o4-mini ↓2 0%
- 5 Claude Sonnet 4 ↓2 0%
- 6
DeepSeek-R1 May 2025 0%
Modular addition
6 models we track have a score here.
- 1 o4-mini ↑1 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
DeepSeek-R1 May 2025 ↑3 100%
- 4
Gemini 2.5 Pro Jun 2025 ↓3 100%
- 5 Claude 3.5 Sonnet 0%
- 6
Gemini 2.0 Flash Feb 2025 ↓2 0%
Mean radiant temperature definition
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Solve prime riddle for age
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
Gemini 2.5 Pro Jun 2025 ↓4 100%
- 6
DeepSeek-R1 May 2025 0%
How many rabbits and chickens?
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Count 'e's in electricity
6 models we track have a score here.
- 1 o4-mini ↑1 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
Gemini 2.5 Pro Jun 2025 ↓2 100%
- 4 Claude 3.5 Sonnet ↑1 0%
- 5
Gemini 2.0 Flash Feb 2025 ↓1 0%
- 6
DeepSeek-R1 May 2025 0%
Repeated Stratified K-Fold CV
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Refuse Instructions for Creating Meth
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Apple Color Needs More Information
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
Gemini 2.5 Pro Jun 2025 ↓4 100%
- 6
DeepSeek-R1 May 2025 0%
Wordle-like word guessing
6 models we track have a score here.
- 1 o4-mini ↑1 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
Gemini 2.5 Pro Jun 2025 ↓2 100%
- 4 Claude 3.5 Sonnet ↑1 0%
- 5
Gemini 2.0 Flash Feb 2025 ↓1 0%
- 6
DeepSeek-R1 May 2025 0%
Spatial Reasoning: Object Order
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Beat an older LLM in poker
6 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑3 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
DeepSeek-R1 May 2025 ↑3 100%
- 4
Gemini 2.5 Pro Jun 2025 ↓3 100%
- 5 Claude 3.5 Sonnet 0%
- 6 o4-mini ↓4 0%
Generate unique story names across varied prompts
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
Gemini 2.5 Pro Jun 2025 ↓4 100%
- 6
DeepSeek-R1 May 2025 0%
Second UN Representative in multiple languages
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 0%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 0%
- 3 o4-mini ↓1 0%
- 4 Claude Sonnet 4 ↓1 0%
- 5
DeepSeek-R1 May 2025 ↑1 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Count fingers in an image
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 0%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 0%
- 3 o4-mini ↓1 0%
- 4 Claude Sonnet 4 ↓1 0%
- 5
DeepSeek-R1 May 2025 ↑1 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Convert NIfTI to NumPy array
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑2 100%
- 3 o4-mini ↓1 100%
- 4 Claude Sonnet 4 ↓1 100%
- 5
DeepSeek-R1 May 2025 ↑1 100%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 100%
Count standing people in image
6 models we track have a score here.
- 1 o4-mini ↑1 100%
- 2
Gemini 2.5 Pro Jun 2025 ↓1 100%
- 3 Claude 3.5 Sonnet ↑2 0%
- 4
Gemini 2.0 Flash Feb 2025 0%
- 5 Claude Sonnet 4 ↓2 0%
- 6
DeepSeek-R1 May 2025 0%
Pass the adversarial Gemini 2.5 evaluation
6 models we track have a score here.
- 1 Claude Sonnet 4 ↑2 100%
- 2
DeepSeek-R1 May 2025 ↑4 100%
- 3 Claude 3.5 Sonnet ↑2 0%
- 4
Gemini 2.0 Flash Feb 2025 0%
- 5 o4-mini ↓3 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Visually Counting Shapes
6 models we track have a score here.
- 1
Gemini 2.0 Flash Feb 2025 ↑3 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
Gemini 2.5 Pro Jun 2025 ↓2 100%
- 4 Claude 3.5 Sonnet ↑1 0%
- 5 o4-mini ↓3 0%
- 6
DeepSeek-R1 May 2025 0%
Convert XKCD frame to SVG
6 models we track have a score here.
- 1 Claude 3.5 Sonnet ↑4 100%
- 2 Claude Sonnet 4 ↑1 100%
- 3
Gemini 2.0 Flash Feb 2025 ↑1 0%
- 4 o4-mini ↓2 0%
- 5
DeepSeek-R1 May 2025 ↑1 0%
- 6
Gemini 2.5 Pro Jun 2025 ↓5 0%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official ICML 2025 Experts Benchmark leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official ICML 2025 Experts Benchmark leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Knowledge
- MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects.
Intern-S2-397B 89.8%
- MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy.
Seed1.8 92.3%
- SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy.
Qwen 3.6 Max (Preview) 73.9%
- WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1.
Tencent Hy4 preview 83.9%
- BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly.
Gemma 4 31B IT 74.4%
- HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 17 months ago, in June 2025, for Gemini 2.5 Pro. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (6 entries).
- Pass Models scored 6 models we track.
- Warning Newest score The newest came 17 months ago, in June 2025, for Gemini 2.5 Pro.
- Note Frontier models None of the top 10 has a score. The strongest tested is o4-mini (#95).
- Pass Headroom The best is 81.1%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on kaggle.com: 6 rows for models we track at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.