RPC-Bench
Models read computer science papers and answer reviewers’ questions. A model judge grades correctness, completeness, and concision against verified review and rebuttal pairs.
- Made by
- Yelin Chen and colleagues
- Released
- April 2026
- Kind
- Research science
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Single turn
- Items from
- Real work and use
- 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
GPT-5 leads with 37.5 among models released by December 2025 GPT-5 leads with 37.5 among models released by December 2025
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. 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 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
- 37.5 GPT-5
- 35.2 GPT-5.2
- 33.4
Gemini 2.5 Pro
- 32.5
Gemini 3 Pro
- 32.0
DeepSeek-V3.2
- 30.8
DeepSeek-V3.1
- 29.2 GLM-4.6V
- 28.8 GLM-4.7
- 27.4 GLM-4.5V
- 25.7 GLM-4.5
- 24.2 Claude Sonnet 4
- 23.3 Qwen3
- 20.1 Claude Sonnet 4.5
- 13.9 HippoRAG2
- 11.0 MemoRAG
- 10.3 VdocRAG
- 9.92 VisRAG
- 8.33 Raptor
- 7.41 Monkey
- 6.96 Docopilot
- 4.27
DocOwl2
- 1 GPT-5 TEXT 37.5
- 2 GPT-5.2 TEXT 35.2
- 3
Gemini 2.5 Pro Jun 2025, TEXT 33.4
- 4
Gemini 3 Pro TEXT 32.5
- 5
DeepSeek-V3.2 TEXT 32.0
- 6
DeepSeek-V3.1 TEXT 30.8
- 7 GLM-4.6V VISUAL 29.2
- 8 GLM-4.7 TEXT 28.8
- 9 GLM-4.5V VISUAL 27.4
- 10 GLM-4.5 TEXT 25.7
- 11 Claude Sonnet 4 TEXT 24.2
- 12 Qwen3 TEXT 23.3
- 13 Claude Sonnet 4.5 TEXT 20.1
- 14 HippoRAG2 TEXT 13.9
- 15 MemoRAG TEXT 11.0
- 16 VdocRAG VISUAL 10.3
- 17 VisRAG VISUAL 9.92
- 18 Raptor TEXT 8.33
- 19 Monkey VISUAL 7.41
- 20 Docopilot VISUAL 6.96
- 21
DocOwl2 VISUAL 4.27
Source: Official RPC-Bench leaderboard
- 1 GPT-5 TEXT 37.5
- 2 GPT-5.2 TEXT 35.2
- 3
Gemini 2.5 Pro Jun 2025, TEXT 33.4
- 4
Gemini 3 Pro TEXT 32.5
- 5
DeepSeek-V3.2 TEXT 32.0
- 6
DeepSeek-V3.1 TEXT 30.8
- 7 GLM-4.6V VISUAL 29.2
- 8 GLM-4.7 TEXT 28.8
- 9 GLM-4.5V VISUAL 27.4
- 10 GLM-4.5 TEXT 25.7
- 11 Claude Sonnet 4 TEXT 24.2
- 12 Claude Sonnet 4.5 TEXT 20.1
- 13 Monkey VISUAL 7.41
- 14
DocOwl2 VISUAL 4.27
Source: Official RPC-Bench leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | GPT-5 | 37.5 | TEXT | Aug 2025 | openai.com | Official leaderboard |
| 2 | GPT-5.2 | 35.2 | TEXT | Dec 2025 | openai.com | Official leaderboard |
| 3 | | 33.4 | TEXT | Jun 2025 | blog.google | Official leaderboard |
| 4 | | 32.5 | TEXT | Nov 2025 | blog.google | Official leaderboard |
| 5 | | 32.0 | TEXT | Dec 2025 | api-docs.deepseek.com | Official leaderboard |
| 6 | | 30.8 | TEXT | Aug 2025 | api-docs.deepseek.com | Official leaderboard |
| 7 | GLM-4.6V | 29.2 | VISUAL | Dec 2025 | github.com | Official leaderboard |
| 8 | GLM-4.7 | 28.8 | TEXT | Dec 2025 | z.ai | Official leaderboard |
| 9 | GLM-4.5V | 27.4 | VISUAL | Aug 2025 | github.com | Official leaderboard |
| 10 | GLM-4.5 | 25.7 | TEXT | Aug 2025 | z.ai | Official leaderboard |
| 11 | Claude Sonnet 4 | 24.2 | TEXT | May 2025 | anthropic.com | Official leaderboard |
| 12 | Qwen3 | 23.3 | TEXT | github.com | Official leaderboard | |
| 13 | Claude Sonnet 4.5 | 20.1 | TEXT | Sep 2025 | anthropic.com | Official leaderboard |
| 14 | HippoRAG2 | 13.9 | TEXT | github.com | Official leaderboard | |
| 15 | MemoRAG | 11.0 | TEXT | github.com | Official leaderboard | |
| 16 | VdocRAG | 10.3 | VISUAL | vdocrag.github.io | Official leaderboard | |
| 17 | VisRAG | 9.92 | VISUAL | github.com | Official leaderboard | |
| 18 | Raptor | 8.33 | TEXT | github.com | Official leaderboard | |
| 19 | Monkey | 7.41 | VISUAL | Nov 2023 | github.com | Official leaderboard |
| 20 | Docopilot | 6.96 | VISUAL | github.com | Official leaderboard | |
| 21 | | 4.27 | VISUAL | Sep 2024 | github.com | Official leaderboard |
| # | Model | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | GPT-5 | 37.5 | TEXT | Aug 2025 | openai.com | Official leaderboard |
| 2 | GPT-5.2 | 35.2 | TEXT | Dec 2025 | openai.com | Official leaderboard |
| 3 | | 33.4 | TEXT | Jun 2025 | blog.google | Official leaderboard |
| 4 | | 32.5 | TEXT | Nov 2025 | blog.google | Official leaderboard |
| 5 | | 32.0 | TEXT | Dec 2025 | api-docs.deepseek.com | Official leaderboard |
| 6 | | 30.8 | TEXT | Aug 2025 | api-docs.deepseek.com | Official leaderboard |
| 7 | GLM-4.6V | 29.2 | VISUAL | Dec 2025 | github.com | Official leaderboard |
| 8 | GLM-4.7 | 28.8 | TEXT | Dec 2025 | z.ai | Official leaderboard |
| 9 | GLM-4.5V | 27.4 | VISUAL | Aug 2025 | github.com | Official leaderboard |
| 10 | GLM-4.5 | 25.7 | TEXT | Aug 2025 | z.ai | Official leaderboard |
| 11 | Claude Sonnet 4 | 24.2 | TEXT | May 2025 | anthropic.com | Official leaderboard |
| 12 | Claude Sonnet 4.5 | 20.1 | TEXT | Sep 2025 | anthropic.com | Official leaderboard |
| 13 | Monkey | 7.41 | VISUAL | Nov 2023 | github.com | Official leaderboard |
| 14 | | 4.27 | VISUAL | Sep 2024 | github.com | Official leaderboard |
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: Official RPC-Bench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-5 | TEXT | 37.5 | |||
| GPT-5.2 | TEXT | 35.2 | |||
| GPT-5 | VISUAL | 35.1 | |||
| TEXT | 33.4 | ||||
| TEXT | 32.5 | ||||
| TEXT | 32.0 | ||||
| GPT-5.2 | VISUAL | 30.9 | |||
| TEXT | 30.8 | ||||
| GLM-4.6V | VISUAL | 29.2 | |||
| GLM-4.7 | TEXT | 28.8 | |||
| GLM-4.5V | VISUAL | 27.4 | |||
| VISUAL | 27.3 | ||||
| GLM-4.5 | TEXT | 25.7 | |||
| VISUAL | 24.3 | ||||
| Claude Sonnet 4 | TEXT | 24.2 | |||
| Claude Sonnet 4.5 | TEXT | 20.1 | |||
| Claude Sonnet 4.5 | VISUAL | 17.5 | |||
| Claude Sonnet 4 | VISUAL | 17.0 | |||
| Monkey | VISUAL | 7.41 | |||
| VISUAL | 4.27 |
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: Official RPC-Bench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-5 | TEXT | 37.5 | |||
| GPT-5.2 | TEXT | 35.2 | |||
| GPT-5 | VISUAL | 35.1 | |||
| TEXT | 33.4 | ||||
| TEXT | 32.5 | ||||
| TEXT | 32.0 | ||||
| GPT-5.2 | VISUAL | 30.9 | |||
| TEXT | 30.8 | ||||
| GLM-4.6V | VISUAL | 29.2 | |||
| GLM-4.7 | TEXT | 28.8 | |||
| GLM-4.5V | VISUAL | 27.4 | |||
| VISUAL | 27.3 | ||||
| GLM-4.5 | TEXT | 25.7 | |||
| VISUAL | 24.3 | ||||
| Claude Sonnet 4 | TEXT | 24.2 | |||
| Claude Sonnet 4.5 | TEXT | 20.1 | |||
| Claude Sonnet 4.5 | VISUAL | 17.5 | |||
| Claude Sonnet 4 | VISUAL | 17.0 | |||
| Monkey | VISUAL | 7.41 | |||
| VISUAL | 4.27 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official RPC-Bench 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 RPC-Bench 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 Science and reasoning
- GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics.
Gemini 4 Carbon 97%
- Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
- BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Fable 5.2 83.7%
- FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics.
Intern-S2-397B 87%
- HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels.
Gemini 3 Pro 78.3%
- LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%
Data audit
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 11 months ago, in December 2025, for GLM-4.7. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (21 entries).
- Pass Models scored 14 models we track, plus 7 other entries on its leaderboard.
- Warning Newest score The newest came 11 months ago, in December 2025, for GLM-4.7.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.2 (#51).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on huggingface.co: 20 rows for models we track and 7 other entries 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.