Benchmark Almanac

EduBench-RU

Gemini 3.1 Pro 3.61 leads since March 2026
Not in model cards

Models plan lessons, explain school subjects, and assist teachers in Russian and Chuvash. A panel of model judges scores responses against a teaching rubric.

Released
March 2026
Kind
Other professions
Level
University
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Capability
Subject
Education
Language
Russian, Chuvash
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Occupations
Educational instruction and library
Strongest model tested
GPT-5.4 #28 on the capability index
Last model added
March 2026
Scores refreshed
November 20th, 2026
29
Entries on its own leaderboard, 21 models we track

Gemini 3.1 Pro leads with 3.61 among models released by March 2026 Gemini 3.1 Pro leads with 3.61 among models released by March 2026

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 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Gemini 3.1 Pro Cloud 3.61
  2. 2 Gemini 3.1 Flash-Lite Cloud 3.40
  3. 3 Claude Opus 4.6 Cloud 3.36
  4. 4 Gemini 2.5 Pro Jun 2025, Cloud 3.34
  5. 5 Claude Sonnet 4.6 Cloud 3.33
  6. 6 DeepSeek-V3.2 Cloud 3.24
  7. 7 Mistral Large 3 Open 3.24
  8. 8 Kimi K2.5 Cloud 3.22
  9. 9 EduLLM-RU 27B Q6_K (ours) Self-hosted 3.21
  10. 10 GLM-5 Cloud 3.20
  11. 11 GPT-5.4 Cloud 3.20
  12. 12 GPT-5.4 Mini Cloud 3.07
  13. 13 Gemini 2.5 Flash Sep 2025, Cloud 3.01
  14. 14 Qwen3.5 27B (base) Open 3.00
  15. 15 Grok 4.1 Fast Cloud 2.94
  16. 16 Qwen3-235B-A22B Open 2.83
  17. 17 EduLLM-RU 32B Q6_K (ours) Self-hosted 2.69
  18. 18 glm-4.7-flash Cloud 2.68
  19. 19 Qwen3-32B Open 2.61
  20. 20 Llama 4 Maverick Open 2.58
  21. 21 YandexGPT 5.1 Pro Russian cloud 2.51
  22. 22 Qwen3-14B Open 2.41
  23. 23 GigaChat-2 Max Russian cloud 2.39
  24. 24 GigaChat-2 Pro Russian cloud 2.36
  25. 25 Qwen3-8B Open 2.34
  26. 26 YandexGPT 5 Lite Russian cloud 2.27
  27. 27 Cohere Command A Cloud 2.23
  28. 28 GigaChat-2 Lite Russian cloud 2.16
  29. 29 Phi-4 Open 1.57

Source: Official EduBench-RU leaderboard

#EntryScoreSettingDateSource
1 Gemini 3.1 Pro 3.61 Cloud Feb 2026 Official leaderboard
2 Gemini 3.1 Flash-Lite 3.40 Cloud Mar 2026 Official leaderboard
3 Claude Opus 4.6 3.36 Cloud Feb 2026 Official leaderboard
4 Gemini 2.5 Pro Jun 2025 3.34 Cloud Jun 2025 Official leaderboard
5 Claude Sonnet 4.6 3.33 Cloud Feb 2026 Official leaderboard
6 DeepSeek-V3.2 3.24 Cloud Dec 2025 Official leaderboard
7 Mistral Large 3 3.24 Open Dec 2025 Official leaderboard
8 Kimi K2.5 3.22 Cloud Jan 2026 Official leaderboard
9 EduLLM-RU 27B Q6_K (ours) 3.21 Self-hosted Official leaderboard
10 GLM-5 3.20 Cloud Feb 2026 Official leaderboard
11 GPT-5.4 3.20 Cloud Mar 2026 Official leaderboard
12 GPT-5.4 Mini 3.07 Cloud Mar 2026 Official leaderboard
13 Gemini 2.5 Flash Sep 2025 3.01 Cloud Sep 2025 Official leaderboard
14 Qwen3.5 27B (base) 3.00 Open Official leaderboard
15 Grok 4.1 Fast 2.94 Cloud Nov 2025 Official leaderboard
16 Qwen3-235B-A22B 2.83 Open Apr 2025 Official leaderboard
17 EduLLM-RU 32B Q6_K (ours) 2.69 Self-hosted Official leaderboard
18 glm-4.7-flash 2.68 Cloud Jan 2026 Official leaderboard
19 Qwen3-32B 2.61 Open Apr 2025 Official leaderboard
20 Llama 4 Maverick 2.58 Open Apr 2025 Official leaderboard
21 YandexGPT 5.1 Pro 2.51 Russian cloud Official leaderboard
22 Qwen3-14B 2.41 Open Apr 2025 Official leaderboard
23 GigaChat-2 Max 2.39 Russian cloud Official leaderboard
24 GigaChat-2 Pro 2.36 Russian cloud Official leaderboard
25 Qwen3-8B 2.34 Open Apr 2025 Official leaderboard
26 YandexGPT 5 Lite 2.27 Russian cloud Official leaderboard
27 Cohere Command A 2.23 Cloud Mar 2025 Official leaderboard
28 GigaChat-2 Lite 2.16 Russian cloud Official leaderboard
29 Phi-4 1.57 Open Dec 2024 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 3.1 Pro 3.61 Cloud Feb 2026 Official leaderboard
2 Gemini 3.1 Flash-Lite 3.40 Cloud Mar 2026 Official leaderboard
3 Claude Opus 4.6 3.36 Cloud Feb 2026 Official leaderboard
4 Gemini 2.5 Pro Jun 2025 3.34 Cloud Jun 2025 Official leaderboard
5 Claude Sonnet 4.6 3.33 Cloud Feb 2026 Official leaderboard
6 DeepSeek-V3.2 3.24 Cloud Dec 2025 Official leaderboard
7 Mistral Large 3 3.24 Open Dec 2025 Official leaderboard
8 Kimi K2.5 3.22 Cloud Jan 2026 Official leaderboard
9 GLM-5 3.20 Cloud Feb 2026 Official leaderboard
10 GPT-5.4 3.20 Cloud Mar 2026 Official leaderboard
11 GPT-5.4 Mini 3.07 Cloud Mar 2026 Official leaderboard
12 Gemini 2.5 Flash Sep 2025 3.01 Cloud Sep 2025 Official leaderboard
13 Grok 4.1 Fast 2.94 Cloud Nov 2025 Official leaderboard
14 Qwen3-235B-A22B 2.83 Open Apr 2025 Official leaderboard
15 glm-4.7-flash 2.68 Cloud Jan 2026 Official leaderboard
16 Qwen3-32B 2.61 Open Apr 2025 Official leaderboard
17 Llama 4 Maverick 2.58 Open Apr 2025 Official leaderboard
18 Qwen3-14B 2.41 Open Apr 2025 Official leaderboard
19 Qwen3-8B 2.34 Open Apr 2025 Official leaderboard
20 Cohere Command A 2.23 Cloud Mar 2025 Official leaderboard
21 Phi-4 1.57 Open Dec 2024 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: EduRU and ChvBn. Each ranks models on one part of the test. Gemini 3.1 Pro leads both. Arrows show how far a model moved from its overall rank among the models we track.

EduRU

21 models we track have a score here.

  1. 1 Gemini 3.1 Pro 3.64
  2. 2 Gemini 3.1 Flash-Lite 3.51
  3. 3 Claude Opus 4.6 3.46
  4. 4 Claude Sonnet 4.6 ↑1 3.43
  5. 5 Gemini 2.5 Pro Jun 2025 ↓1 3.38
  6. 6 DeepSeek-V3.2 3.37
  7. 7 Mistral Large 3 3.35
  8. 8 Kimi K2.5 3.35
  9. 9 GLM-5 3.34
  10. 10 GPT-5.4 3.33

Number one over time

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

Google DeepMind
Own leaderboard Not tested since Mar 2026

Source: Official EduBench-RU 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
1.50 2.00 2.50 3.00 3.50 4.00 2025 2026 1.50 2.00 2.50 3.00 3.50 4.00 2025 2026

Source: Official EduBench-RU 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 Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 8 months ago, in March 2026, for GPT-5.4 Mini. 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 (29 entries).
  • Pass Models scored 21 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 8 months ago, in March 2026, for GPT-5.4 Mini.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.4 (#28).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 21 rows for models we track and 8 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.

How we check