Benchmark Almanac

Compass Academic Leaderboard

DeepSeek-V3 72.6% leads since December 2024
Not in model cards

Models answer academic reasoning tasks. OpenCompass publishes model rankings and predictions across the underlying evaluations.

The leaderboard aggregates results from component evaluations.

Made by
OpenCompass
Released
December 2024
Kind
Composite indices
Level
University
Scored
Settled by events
Answers
Short answer
Tasks
All public
Measures
Capability
In and out
Text → Text
Interaction
Single turn
Items from
Another benchmark
Refresh
One fixed set
Run by
Anyone
Strongest model tested
DeepSeek-V3 #155 on the capability index
Last model added
March 2025
Scores refreshed
November 20th, 2026
16
Entries on its own leaderboard, 15 models we track

DeepSeek-V3 leads with 72.6% among models released by March 2025 DeepSeek-V3 leads with 72.6% among models released by March 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. 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 March 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 DeepSeek-V3 72.6%
  2. 2 Qwen2.5-Max 71.8%
  3. 3 Qwen2.5-72B 68.5%
  4. 4 Qwen2.5-32B 68%
  5. 5 DeepSeek-V2.5-1210 67.6%
  6. 6 Llama 3.3 70B Instruct 67.2%
  7. 7 MiniMax-Text-01 67.1%
  8. 8 DeepSeek-V2.5 Sep 2024 64.5%
  9. 9 Qwen2.5-14B 64.3%
  10. 10 Gemma 3 27B 64.3%
  11. 11 Qwen2.5-7B 56.5%
  12. 12 Gemma 2 27B 56%
  13. 13 InternLM2.5-Chat-7B 51%
  14. 14 Llama 3.1-8B 50.7%
  15. 15 Gemma 2 9B 50.3%
  16. 16 Llama 3.2 3B 42.9%

Source: Official Compass Academic Leaderboard leaderboard

#EntryScoreSettingDateSource
1 DeepSeek-V3 72.6% Dec 2024 Official leaderboard
2 Qwen2.5-Max 71.8% Jan 2025 Official leaderboard
3 Qwen2.5-72B 68.5% Sep 2024 Official leaderboard
4 Qwen2.5-32B 68% Sep 2024 Official leaderboard
5 DeepSeek-V2.5-1210 67.6% Dec 2024 Official leaderboard
6 Llama 3.3 70B Instruct 67.2% Dec 2024 Official leaderboard
7 MiniMax-Text-01 67.1% Jan 2025 Official leaderboard
8 DeepSeek-V2.5 Sep 2024 64.5% Sep 2024 Official leaderboard
9 Qwen2.5-14B 64.3% Sep 2024 Official leaderboard
10 Gemma 3 27B 64.3% Mar 2025 Official leaderboard
11 Qwen2.5-7B 56.5% Sep 2024 Official leaderboard
12 Gemma 2 27B 56% Jun 2024 Official leaderboard
13 InternLM2.5-Chat-7B open weights, 7B 51% Official leaderboard
14 Llama 3.1-8B 50.7% Jul 2024 Official leaderboard
15 Gemma 2 9B 50.3% Jun 2024 Official leaderboard
16 Llama 3.2 3B 42.9% Sep 2024 Official leaderboard
#ModelScoreSettingDateSource
1 DeepSeek-V3 72.6% Dec 2024 Official leaderboard
2 Qwen2.5-Max 71.8% Jan 2025 Official leaderboard
3 Qwen2.5-72B 68.5% Sep 2024 Official leaderboard
4 Qwen2.5-32B 68% Sep 2024 Official leaderboard
5 DeepSeek-V2.5-1210 67.6% Dec 2024 Official leaderboard
6 Llama 3.3 70B Instruct 67.2% Dec 2024 Official leaderboard
7 MiniMax-Text-01 67.1% Jan 2025 Official leaderboard
8 DeepSeek-V2.5 Sep 2024 64.5% Sep 2024 Official leaderboard
9 Qwen2.5-14B 64.3% Sep 2024 Official leaderboard
10 Gemma 3 27B 64.3% Mar 2025 Official leaderboard
11 Qwen2.5-7B 56.5% Sep 2024 Official leaderboard
12 Gemma 2 27B 56% Jun 2024 Official leaderboard
13 Llama 3.1-8B 50.7% Jul 2024 Official leaderboard
14 Gemma 2 9B 50.3% Jun 2024 Official leaderboard
15 Llama 3.2 3B 42.9% Sep 2024 Official leaderboard

Rankings by part

We read 15 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as IFEval, BBH and GPQA_diamond. Qwen2.5-Max leads 6 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.

IFEval

15 models we track have a score here.

  1. 1 Llama 3.3 70B Instruct ↑5 88.5%
  2. 2 DeepSeek-V3 ↓1 85.8%
  3. 3 Qwen2.5-Max ↓1 84.7%
  4. 4 MiniMax-Text-01 ↑3 83.5%
  5. 5 Qwen2.5-72B ↓2 83%
  6. 6 Gemma 3 27B ↑4 81%
  7. 7 DeepSeek-V2.5-1210 ↓2 79.3%
  8. 8 Qwen2.5-32B ↓4 78.9%
  9. 9 Gemma 2 27B ↑3 78.2%
  10. 10 DeepSeek-V2.5 Sep 2024 ↓2 77.6%

2 models have held number one on the official leaderboard since its release in December 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen DeepSeek
Own leaderboard Not tested since Mar 2025
DeepSeek-V3 72.6%

Source: Official Compass Academic Leaderboard leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) DeepSeek-V3 Qwen2.5-72B Gemma 2 27B 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) DeepSeek-V3 DeepSeek-V2.5

Source: Official Compass Academic Leaderboard 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 Other

All 69 in Other
  • Artificial Analysis Intelligence Index v5 In use Version 5 of Artificial Analysis's index: nine evaluations in four equal categories, most with private test sets, run three times each in its own agent harness. Claude Fable 5.2 64.6
  • Personal Superintelligence Index (Meta) In use Meta's composite of seven personal and knowledge-work evals: Agents' Last Exam, AutomationBench-AA, FocusBench, JobBench, GDP.pdf, VBT and DRACO.
  • Vals Science Index In use Agents are scored on the equal average of BioMysteryBench, MysteryMechanism and Terminal-Bench-Science. GPT-6.2 Astra 68.4%
  • Decision Index Public v0.3 In use Decision models return probabilities over labeled options in public tasks covering knowledge, language, retrieval, tools, and arts. A chance-corrected weighted aggregate combines the task scores. Granite Switch 4.1 30B 51.3
  • SAGE v1.1 (Vals) In use Models grade handwritten undergraduate math and computer science work against rubrics. Scored by class-balanced accuracy of each rubric check against expert grading. Claude Opus 5.5 91.3%
  • MTEB Fading Massive Text Embedding Benchmark: retrieval, clustering, classification across languages and modalities. Qwen3-Embedding-8B 72.2%

Data audit

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 20 months ago, in March 2025, for Gemma 3 27B. 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 (16 entries).
  • Pass Models scored 15 models we track, plus 1 other entry on its leaderboard.
  • Warning Newest score The newest came 20 months ago, in March 2025, for Gemma 3 27B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is DeepSeek-V3 (#155).
  • Pass Headroom The best is 72.6%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on huggingface.co: 15 rows for models we track and 1 other entry 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.

How we check