ResearchCodeBench
Agents read machine learning papers and fill missing Python code in the accompanying implementations. Unit tests check each of 212 snippets from 20 papers.
The comparison available November 19, 2026 tests GPT-4o mini and GPT-4.1 mini with Codex. GPT-4.1 mini led at 41.01% originally and 40.74% on Harbor.
- Made by
- ResearchCodeBench team
- Released
- May 2025
- Kind
- Machine learning engineering
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Terminal
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Agents
- Strongest model tested
- o3 #84 on the capability index
- Last model added
- September 2025
- Scores refreshed
- November 21st, 2026
Gemini 2.5 Pro leads with 64.2% among models released by September 2025 Gemini 2.5 Pro leads with 64.2% among models released by September 2025
Best score per entry from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in September 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
- 64.2
Gemini 2.5 Pro
- 59.4 o3
- 58.5 o4-mini
- 52.4 o3-mini
- 51.9 Claude 3.7 Sonnet
- 50.0 GPT-4.1
- 48.6 Claude-3.5-Sonnet (snapshot unspecified)
- 48.1 o1
- 45.8 DeepSeek-R1 (snapshot unspecified)
- 45.3
Gemini 2.5 Flash
- 42.9 Grok 3
- 42.5 DeepSeek-Chat-V3-0324
- 42.5 GPT-4.1 mini
- 41.0 GPT-4o (snapshot unspecified)
- 37.7 Claude 3.5 Haiku
- 37.3
Gemini 2.0 Flash
- 35.4
Mistral Medium 3
- 33.5 Mistral-Codestral-2501
- 31.1 Command-A
- 30.7
Gemini 2.0 Flash-Lite
- 27.4
Llama 4 Maverick
- 25.9
Qwen2.5-Coder-32B
- 25.0 Nova-Pro-V1
- 23.1 GPT-4o-Mini (snapshot unspecified)
- 22.2 Grok-2
- 19.8 Grok-3 mini
- 18.4
Llama 4 Scout
- 15.1 GPT-4.1 nano
- 12.3
Llama 3.3 70B
- 8.02
Qwen2.5-Turbo
- 1.42 Mistral-Codestral-Mamba
- 1
Gemini 2.5 Pro Jun 2025 64.2%
- 2 o3 high 59.4%
- 3 o4-mini high 58.5%
- 4 o3-mini high 52.4%
- 5 Claude 3.7 Sonnet 51.9%
- 6 GPT-4.1 50%
- 7 Claude-3.5-Sonnet (snapshot unspecified) 48.6%
- 8 o1 high 48.1%
- 9 DeepSeek-R1 (snapshot unspecified) 45.8%
- 10
Gemini 2.5 Flash Sep 2025 45.3%
- 11 Grok 3 42.9%
- 12 DeepSeek-Chat-V3-0324 42.5%
- 13 GPT-4.1 mini 42.5%
- 14 GPT-4o (snapshot unspecified) 41%
- 15 Claude 3.5 Haiku 37.7%
- 16
Gemini 2.0 Flash Feb 2025 37.3%
- 17
Mistral Medium 3 35.4%
- 18 Mistral-Codestral-2501 33.5%
- 19 Command-A 31.1%
- 20
Gemini 2.0 Flash-Lite 30.7%
- 21
Llama 4 Maverick 27.4%
- 22
Qwen2.5-Coder-32B 25.9%
- 23 Nova-Pro-V1 25%
- 24 GPT-4o-Mini (snapshot unspecified) 23.1%
- 25 Grok-2 Dec 2024 22.2%
- 26 Grok-3 mini high 19.8%
- 27
Llama 4 Scout 18.4%
- 28 GPT-4.1 nano 15.1%
- 29
Llama 3.3 70B 12.3%
- 30
Qwen2.5-Turbo 8.02%
- 31 Mistral-Codestral-Mamba 1.42%
Source: Official ResearchCodeBench leaderboard
- 64.2
Gemini 2.5 Pro
- 59.4 o3
- 58.5 o4-mini
- 52.4 o3-mini
- 51.9 Claude 3.7 Sonnet
- 50.0 GPT-4.1
- 48.1 o1
- 45.3
Gemini 2.5 Flash
- 42.9 Grok 3
- 42.5 GPT-4.1 mini
- 37.7 Claude 3.5 Haiku
- 37.3
Gemini 2.0 Flash
- 35.4
Mistral Medium 3
- 30.7
Gemini 2.0 Flash-Lite
- 27.4
Llama 4 Maverick
- 25.9
Qwen2.5-Coder-32B
- 22.2 Grok-2
- 19.8 Grok-3 mini
- 18.4
Llama 4 Scout
- 15.1 GPT-4.1 nano
- 12.3
Llama 3.3 70B
- 8.02
Qwen2.5-Turbo
- 1
Gemini 2.5 Pro Jun 2025 64.2%
- 2 o3 high 59.4%
- 3 o4-mini high 58.5%
- 4 o3-mini high 52.4%
- 5 Claude 3.7 Sonnet 51.9%
- 6 GPT-4.1 50%
- 7 o1 high 48.1%
- 8
Gemini 2.5 Flash Sep 2025 45.3%
- 9 Grok 3 42.9%
- 10 GPT-4.1 mini 42.5%
- 11 Claude 3.5 Haiku 37.7%
- 12
Gemini 2.0 Flash Feb 2025 37.3%
- 13
Mistral Medium 3 35.4%
- 14
Gemini 2.0 Flash-Lite 30.7%
- 15
Llama 4 Maverick 27.4%
- 16
Qwen2.5-Coder-32B 25.9%
- 17 Grok-2 Dec 2024 22.2%
- 18 Grok-3 mini high 19.8%
- 19
Llama 4 Scout 18.4%
- 20 GPT-4.1 nano 15.1%
- 21
Llama 3.3 70B 12.3%
- 22
Qwen2.5-Turbo 8.02%
Source: Official ResearchCodeBench leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 64.2% | Jun 2025 | Official leaderboard | |
| 2 | o3 | 59.4% | high | Apr 2025 | Official leaderboard |
| 3 | o4-mini | 58.5% | high | Apr 2025 | Official leaderboard |
| 4 | o3-mini | 52.4% | high | Jan 2025 | Official leaderboard |
| 5 | Claude 3.7 Sonnet | 51.9% | Feb 2025 | Official leaderboard | |
| 6 | GPT-4.1 | 50% | Apr 2025 | Official leaderboard | |
| 7 | Claude-3.5-Sonnet (snapshot unspecified) | 48.6% | Official leaderboard | ||
| 8 | o1 | 48.1% | high | Dec 2024 | Official leaderboard |
| 9 | DeepSeek-R1 (snapshot unspecified) | 45.8% | Official leaderboard | ||
| 10 | | 45.3% | Sep 2025 | Official leaderboard | |
| 11 | Grok 3 | 42.9% | Apr 2025 | Official leaderboard | |
| 12 | DeepSeek-Chat-V3-0324 | 42.5% | Official leaderboard | ||
| 13 | GPT-4.1 mini | 42.5% | Apr 2025 | Official leaderboard | |
| 14 | GPT-4o (snapshot unspecified) | 41% | Official leaderboard | ||
| 15 | Claude 3.5 Haiku | 37.7% | Oct 2024 | Official leaderboard | |
| 16 | | 37.3% | Feb 2025 | Official leaderboard | |
| 17 | | 35.4% | May 2025 | Official leaderboard | |
| 18 | Mistral-Codestral-2501 | 33.5% | Official leaderboard | ||
| 19 | Command-A | 31.1% | Official leaderboard | ||
| 20 | | 30.7% | Feb 2025 | Official leaderboard | |
| 21 | | 27.4% | Apr 2025 | Official leaderboard | |
| 22 | | 25.9% | Sep 2024 | Official leaderboard | |
| 23 | Nova-Pro-V1 | 25% | Official leaderboard | ||
| 24 | GPT-4o-Mini (snapshot unspecified) | 23.1% | Official leaderboard | ||
| 25 | Grok-2 Dec 2024 | 22.2% | Dec 2024 | Official leaderboard | |
| 26 | Grok-3 mini | 19.8% | high | Apr 2025 | Official leaderboard |
| 27 | | 18.4% | Apr 2025 | Official leaderboard | |
| 28 | GPT-4.1 nano | 15.1% | Apr 2025 | Official leaderboard | |
| 29 | | 12.3% | Dec 2024 | Official leaderboard | |
| 30 | | 8.02% | Nov 2024 | Official leaderboard | |
| 31 | Mistral-Codestral-Mamba | 1.42% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 64.2% | Jun 2025 | Official leaderboard | |
| 2 | o3 | 59.4% | high | Apr 2025 | Official leaderboard |
| 3 | o4-mini | 58.5% | high | Apr 2025 | Official leaderboard |
| 4 | o3-mini | 52.4% | high | Jan 2025 | Official leaderboard |
| 5 | Claude 3.7 Sonnet | 51.9% | Feb 2025 | Official leaderboard | |
| 6 | GPT-4.1 | 50% | Apr 2025 | Official leaderboard | |
| 7 | o1 | 48.1% | high | Dec 2024 | Official leaderboard |
| 8 | | 45.3% | Sep 2025 | Official leaderboard | |
| 9 | Grok 3 | 42.9% | Apr 2025 | Official leaderboard | |
| 10 | GPT-4.1 mini | 42.5% | Apr 2025 | Official leaderboard | |
| 11 | Claude 3.5 Haiku | 37.7% | Oct 2024 | Official leaderboard | |
| 12 | | 37.3% | Feb 2025 | Official leaderboard | |
| 13 | | 35.4% | May 2025 | Official leaderboard | |
| 14 | | 30.7% | Feb 2025 | Official leaderboard | |
| 15 | | 27.4% | Apr 2025 | Official leaderboard | |
| 16 | | 25.9% | Sep 2024 | Official leaderboard | |
| 17 | Grok-2 Dec 2024 | 22.2% | Dec 2024 | Official leaderboard | |
| 18 | Grok-3 mini | 19.8% | high | Apr 2025 | Official leaderboard |
| 19 | | 18.4% | Apr 2025 | Official leaderboard | |
| 20 | GPT-4.1 nano | 15.1% | Apr 2025 | Official leaderboard | |
| 21 | | 12.3% | Dec 2024 | Official leaderboard | |
| 22 | | 8.02% | Nov 2024 | Official leaderboard |
2 models have held number one on the official leaderboard since its release in May 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official ResearchCodeBench 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 ResearchCodeBench 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 AI research
- PaperBench In use Agent replicates ICML 2024 papers from scratch. LLM judge grades against author-written hierarchical rubrics.
Qwen 3.8 Max 93%
- MLS-Bench In use Agent must invent one algorithmic change in an ML research scaffold whose gain transfers across settings. Scored relative to human baselines. Claude Fable 5.1 50.3%
- PostTrainBench 1.2 In use CLI agent gets a base model, one H100 and a time budget to post-train it for a target benchmark. Scored on normalized gain. Claude Fable 5.1 44.6%
- MLE-bench In use Kaggle-style ML engineering competitions. Agent trains models and is scored against competition leaderboards. GPT-6.2 Astra 96.1%
- PMPP-Hard In use Agents write GPU kernels for 69 tasks that must be correct, and on 48 of them also run within 1.25 times the reference time. The score is the area under the solved-tasks curve over a 10-hour budget, out of 690 task-hours. Kimi K3 0.71
- KernelBench In use Replace PyTorch operators with faster custom GPU kernels. Scored on correctness and speedup.
MiniMax-M3 28.8%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash. 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 (31 entries).
- Pass Models scored 22 models we track, plus 9 other entries on its leaderboard.
- Warning Newest score The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is o3 (#84).
- Pass Headroom The best is 64.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 23 rows for models we track and 9 other entries at the last read.
- Pass Last read November 21st, 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.