ProOPF-B
Models translate operational requirements into executable optimal power flow programs. They change grid parameters, objectives, and constraints. The evaluator checks implementations against expert references.
Grid Studio scored 19 of 121 test cases and restored omitted specifications from six source records. ProOPF-D is the training dataset.
- Made by
- Zhejiang University, Peking University, Xi’an Jiaotong University, and China Southern Power Grid
- Released
- February 2026
- Kind
- Scientific code
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Environment
- Code sandbox
- Items from
- Written by experts
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5.2 #51 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 20th, 2026
Gemini 3 Pro leads with 37.2% among models released by December 2025
Best score per model from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.
The newest model with a score here came out in December 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 3 Pro Few-shot 37.2%
- 2
Qwen3-30B-A3B SFT (ProOPF-D), few-shot 35.5%
- 3 Claude Sonnet 4.5 Few-shot 33.9%
- 4
DeepSeek-V3.2 Few-shot 29.8%
- 5
Qwen3-Coder-480B-A35B Few-shot 24%
- 6 GPT-5.2 Few-shot 14.1%
Source: The ProOPF-B paper
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 37.2% | Few-shot | Nov 2025 | Official leaderboard |
| 2 | | 35.5% | SFT (ProOPF-D), few-shot | Apr 2025 | Official leaderboard |
| 3 | Claude Sonnet 4.5 | 33.9% | Few-shot | Sep 2025 | Official leaderboard |
| 4 | | 29.8% | Few-shot | Dec 2025 | Official leaderboard |
| 5 | | 24% | Few-shot | Jul 2025 | Official leaderboard |
| 6 | GPT-5.2 | 14.1% | Few-shot | Dec 2025 | Official leaderboard |
Rankings by part
We read 4 parts from the benchmark’s paper, each ranking models on one part of the test, such as Level 1, Level 2 and Level 3. Qwen3-30B-A3B leads 2 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Level 1
6 models we track have a score here.
- 1
Gemini 3 Pro 94.4%
- 2
DeepSeek-V3.2 ↑2 94.4%
- 3
Qwen3-Coder-480B-A35B ↑2 80.6%
- 4 Claude Sonnet 4.5 ↓1 77.8%
- 5
Qwen3-30B-A3B ↓3 75%
- 6 GPT-5.2 33.3%
Level 2
6 models we track have a score here.
- 1
Qwen3-30B-A3B ↑1 33.3%
- 2 Claude Sonnet 4.5 ↑1 6.67%
- 3
Gemini 3 Pro ↓2 6.67%
- 4 GPT-5.2 ↑2 6.67%
- 5
Qwen3-Coder-480B-A35B 0%
- 6
DeepSeek-V3.2 ↓2 0%
Level 3
6 models we track have a score here.
- 1 Claude Sonnet 4.5 ↑2 37.9%
- 2
Gemini 3 Pro ↓1 31%
- 3
Qwen3-30B-A3B ↓1 20.7%
- 4 GPT-5.2 ↑2 10.3%
- 5
DeepSeek-V3.2 ↓1 6.90%
- 6
Qwen3-Coder-480B-A35B ↓1 5.56%
Level 4
6 models we track have a score here.
- 1
Qwen3-30B-A3B ↑1 11.5%
- 2
Qwen3-Coder-480B-A35B ↑3 0%
- 3 Claude Sonnet 4.5 0%
- 4
Gemini 3 Pro ↓3 0%
- 5
DeepSeek-V3.2 ↓1 0%
- 6 GPT-5.2 0%
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The ProOPF-B paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Few-shot | 37.2% | ||||
| SFT (ProOPF-D), few-shot | 35.5% | ||||
| Claude Sonnet 4.5 | Few-shot | 33.9% | |||
| Few-shot | 29.8% | ||||
| SFT (ProOPF-D), zero-shot | 27.3% | ||||
| Zero-shot | 26.5% | ||||
| Zero-shot | 26.4% | ||||
| Few-shot | 24% | ||||
| Claude Sonnet 4.5 | Zero-shot | 22.8% | |||
| Few-shot | 14.9% | ||||
| GPT-5.2 | Few-shot | 14.1% | |||
| GPT-5.2 | Zero-shot | 10.4% | |||
| Zero-shot | 5.46% | ||||
| Zero-shot | 0% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The ProOPF-B paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official ProOPF-B 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 Coding
- SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
- LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests.
DeepSeek-V4-Pro 93.5%
- SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
- DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
- SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases.
Seed2.1 Pro 59.8%
- Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance.
DeepSeek V4.1 Flash 3,471
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 11 months ago, in December 2025, for GPT-5.2. 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 benchmark’s paper (6 entries).
- Pass Models scored 6 models we track.
- Warning Newest score The newest came 11 months ago, in December 2025, for GPT-5.2.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.2 (#51).
- Pass Headroom The best is 37.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 14 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.