BhashaBench-Krishi
Version V1
Models answer English and Hindi questions from Indian government agricultural exams. Topics include agronomy, horticulture and animal husbandry.
The test contains 15,405 questions. The authors require approval to download it.
- Made by
- BharatGen
- Released
- October 2025
- Kind
- Exam questions
- Level
- University
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Interaction
- Single turn
- Items from
- Exams and contests
- Refresh
- New versions replace it
- Run by
- Anyone
- Strongest model tested
- gpt-oss-120b #126 on the capability index
- Last model added
- August 2025
- Scores refreshed
- November 20th, 2026
gpt-4o (snapshot unspecified) tops the leaderboard with 73.5% Qwen3-235B-A22B-Instruct leads with 72.7% among models released by August 2025
Best score per entry 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. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Qwen3-235B-A22B-Instruct, at 72.7%. This view shows only models we track, each at its best.
The newest model with a score here came out in August 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
- 73.5 gpt-4o (snapshot unspecified)
- 72.7
Qwen3-235B-A22B-Instruct
- 69.4 gpt-oss-120b
- 59.7 deepseek-v3 (snapshot unspecified)
- 58.1
Gemma 2 27B
- 53.2
Gemma 2 9B
- 52.5 gemma-2-27b-it
- 46.7 gemma-2-9b-it
- 45.0 Llama-3.1-8B-Instruct
- 44.9 Pangea-7B
- 44.7 gpt-oss-20b
- 42.5
Qwen2.5-3B
- 40.7 Indic-gemma-7b-finetuned-sft-Navarasa-2.0
- 39.9 Qwen2.5-3B-Instruct
- 38.8 gemma-2-2b
- 38.5 Llama-3.2-3B-Instruct
- 38.1
Llama 3.1-8B
- 36.6 Nemotron-4-Mini-Hindi-4B-Base
- 36.4 aya-23-8B
- 36.0 Granite-3.1-2B-Instruct
- 35.7 Nemotron-4-Mini-Hindi-4B-Instruct
- 34.5
Gemma 2 2B IT
- 32.2 Param-1
- 32.2 Granite-3.1-3B-A800M-Base
- 32.0
Llama 3.2 3B
- 30.2 sarvam-1
- 28.9
Llama 3.2 1B
- 28.6 Llama-3.2-1B-Instruct
- 27.7 sarvam-2b-v0.5
- 27.1
Gemma 3 270M
- 26.2
Gemma 3 270M Base
- 1 gpt-4o (snapshot unspecified) Zero-shot 73.5%
- 2
Qwen3-235B-A22B-Instruct Jul 2025, Zero-shot 72.7%
- 3 gpt-oss-120b Zero-shot 69.4%
- 4 deepseek-v3 (snapshot unspecified) Zero-shot 59.7%
- 5
Gemma 2 27B Zero-shot 58.1%
- 6
Gemma 2 9B Zero-shot 53.2%
- 7 gemma-2-27b-it Zero-shot 52.5%
- 8 gemma-2-9b-it Zero-shot 46.7%
- 9 Llama-3.1-8B-Instruct Zero-shot 45%
- 10 Pangea-7B Zero-shot 44.9%
- 11 gpt-oss-20b Zero-shot 44.7%
- 12
Qwen2.5-3B Zero-shot 42.5%
- 13 Indic-gemma-7b-finetuned-sft-Navarasa-2.0 Zero-shot 40.7%
- 14 Qwen2.5-3B-Instruct Zero-shot 39.9%
- 15 gemma-2-2b Zero-shot 38.8%
- 16 Llama-3.2-3B-Instruct Zero-shot 38.5%
- 17
Llama 3.1-8B Zero-shot 38.1%
- 18 Nemotron-4-Mini-Hindi-4B-Base Zero-shot 36.6%
- 19 aya-23-8B Zero-shot 36.4%
- 20 Granite-3.1-2B-Instruct Zero-shot 36%
- 21 Nemotron-4-Mini-Hindi-4B-Instruct Zero-shot 35.7%
- 22
Gemma 2 2B IT Zero-shot 34.5%
- 23 Param-1 Zero-shot 32.2%
- 24 Granite-3.1-3B-A800M-Base Zero-shot 32.2%
- 25
Llama 3.2 3B Zero-shot 32%
- 26 sarvam-1 Zero-shot 30.2%
- 27
Llama 3.2 1B Zero-shot 28.9%
- 28 Llama-3.2-1B-Instruct Zero-shot 28.6%
- 29 sarvam-2b-v0.5 Zero-shot 27.7%
- 30
Gemma 3 270M Zero-shot 27.1%
- 31
Gemma 3 270M Base Zero-shot 26.2%
Source: The BhashaBench-Krishi paper
- 1
Qwen3-235B-A22B-Instruct Jul 2025, Zero-shot 72.7%
- 2 gpt-oss-120b Zero-shot 69.4%
- 3
Gemma 2 27B Zero-shot 58.1%
- 4
Gemma 2 9B Zero-shot 53.2%
- 5 gpt-oss-20b Zero-shot 44.7%
- 6
Qwen2.5-3B Zero-shot 42.5%
- 7
Llama 3.1-8B Zero-shot 38.1%
- 8 Granite-3.1-2B-Instruct Zero-shot 36%
- 9
Gemma 2 2B IT Zero-shot 34.5%
- 10 Granite-3.1-3B-A800M-Base Zero-shot 32.2%
- 11
Llama 3.2 3B Zero-shot 32%
- 12
Llama 3.2 1B Zero-shot 28.9%
- 13
Gemma 3 270M Zero-shot 27.1%
- 14
Gemma 3 270M Base Zero-shot 26.2%
Source: The BhashaBench-Krishi paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | gpt-4o (snapshot unspecified) | 73.5% | Zero-shot | Official leaderboard | |
| 2 | | 72.7% | Zero-shot | Jul 2025 | Official leaderboard |
| 3 | gpt-oss-120b | 69.4% | Zero-shot | Aug 2025 | Official leaderboard |
| 4 | deepseek-v3 (snapshot unspecified) | 59.7% | Zero-shot | Official leaderboard | |
| 5 | | 58.1% | Zero-shot | Jun 2024 | Official leaderboard |
| 6 | | 53.2% | Zero-shot | Jun 2024 | Official leaderboard |
| 7 | gemma-2-27b-it | 52.5% | Zero-shot | Official leaderboard | |
| 8 | gemma-2-9b-it | 46.7% | Zero-shot | Official leaderboard | |
| 9 | Llama-3.1-8B-Instruct | 45% | Zero-shot | Official leaderboard | |
| 10 | Pangea-7B | 44.9% | Zero-shot | Official leaderboard | |
| 11 | gpt-oss-20b | 44.7% | Zero-shot | Aug 2025 | Official leaderboard |
| 12 | | 42.5% | Zero-shot | Sep 2024 | Official leaderboard |
| 13 | Indic-gemma-7b-finetuned-sft-Navarasa-2.0 | 40.7% | Zero-shot | Official leaderboard | |
| 14 | Qwen2.5-3B-Instruct | 39.9% | Zero-shot | Official leaderboard | |
| 15 | gemma-2-2b | 38.8% | Zero-shot | Official leaderboard | |
| 16 | Llama-3.2-3B-Instruct | 38.5% | Zero-shot | Official leaderboard | |
| 17 | | 38.1% | Zero-shot | Jul 2024 | Official leaderboard |
| 18 | Nemotron-4-Mini-Hindi-4B-Base | 36.6% | Zero-shot | Official leaderboard | |
| 19 | aya-23-8B | 36.4% | Zero-shot | Official leaderboard | |
| 20 | Granite-3.1-2B-Instruct | 36% | Zero-shot | Dec 2024 | Official leaderboard |
| 21 | Nemotron-4-Mini-Hindi-4B-Instruct | 35.7% | Zero-shot | Official leaderboard | |
| 22 | | 34.5% | Zero-shot | Jul 2024 | Official leaderboard |
| 23 | Param-1 | 32.2% | Zero-shot | Official leaderboard | |
| 24 | Granite-3.1-3B-A800M-Base | 32.2% | Zero-shot | Dec 2024 | Official leaderboard |
| 25 | | 32% | Zero-shot | Sep 2024 | Official leaderboard |
| 26 | sarvam-1 | 30.2% | Zero-shot | Official leaderboard | |
| 27 | | 28.9% | Zero-shot | Sep 2024 | Official leaderboard |
| 28 | Llama-3.2-1B-Instruct | 28.6% | Zero-shot | Official leaderboard | |
| 29 | sarvam-2b-v0.5 | 27.7% | Zero-shot | Official leaderboard | |
| 30 | | 27.1% | Zero-shot | Aug 2025 | Official leaderboard |
| 31 | | 26.2% | Zero-shot | Aug 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 72.7% | Zero-shot | Jul 2025 | Official leaderboard |
| 2 | gpt-oss-120b | 69.4% | Zero-shot | Aug 2025 | Official leaderboard |
| 3 | | 58.1% | Zero-shot | Jun 2024 | Official leaderboard |
| 4 | | 53.2% | Zero-shot | Jun 2024 | Official leaderboard |
| 5 | gpt-oss-20b | 44.7% | Zero-shot | Aug 2025 | Official leaderboard |
| 6 | | 42.5% | Zero-shot | Sep 2024 | Official leaderboard |
| 7 | | 38.1% | Zero-shot | Jul 2024 | Official leaderboard |
| 8 | Granite-3.1-2B-Instruct | 36% | Zero-shot | Dec 2024 | Official leaderboard |
| 9 | | 34.5% | Zero-shot | Jul 2024 | Official leaderboard |
| 10 | Granite-3.1-3B-A800M-Base | 32.2% | Zero-shot | Dec 2024 | Official leaderboard |
| 11 | | 32% | Zero-shot | Sep 2024 | Official leaderboard |
| 12 | | 28.9% | Zero-shot | Sep 2024 | Official leaderboard |
| 13 | | 27.1% | Zero-shot | Aug 2025 | Official leaderboard |
| 14 | | 26.2% | Zero-shot | Aug 2025 | Official leaderboard |
Rankings by part
We read 10 parts from the benchmark’s paper, each ranking models on one part of the test, such as English, Hindi and Easy. Qwen3-235B-A22B-Instruct leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
English
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 74.6%
- 2 gpt-oss-120b 71.4%
- 3
Gemma 2 27B 59.8%
- 4
Gemma 2 9B 55.2%
- 5 gpt-oss-20b 46.6%
- 6
Qwen2.5-3B 44.6%
- 7
Llama 3.1-8B 39.5%
- 8 Granite-3.1-2B-Instruct 37.7%
- 9
Gemma 2 2B IT 35.9%
- 10 Granite-3.1-3B-A800M-Base 33.4%
Hindi
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 64.1%
- 2 gpt-oss-120b 60.3%
- 3
Gemma 2 27B 50.4%
- 4
Gemma 2 9B 43.9%
- 5 gpt-oss-20b 36.3%
- 6
Qwen2.5-3B 32.7%
- 7
Llama 3.1-8B 31.4%
- 8
Llama 3.2 3B ↑3 28.7%
- 9 Granite-3.1-2B-Instruct ↓1 27.9%
- 10
Gemma 2 2B IT ↓1 27.7%
Easy
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 78.3%
- 2 gpt-oss-120b 74.9%
- 3
Gemma 2 27B 69.3%
- 4
Gemma 2 9B 64.8%
- 5 gpt-oss-20b 53.4%
- 6
Qwen2.5-3B 50.3%
- 7
Llama 3.1-8B 44%
- 8 Granite-3.1-2B-Instruct 40%
- 9
Gemma 2 2B IT 38%
- 10
Llama 3.2 3B ↑1 36.4%
Medium
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 69.8%
- 2 gpt-oss-120b 65.9%
- 3
Gemma 2 27B 51.5%
- 4
Gemma 2 9B 46.3%
- 5 gpt-oss-20b 39.6%
- 6
Qwen2.5-3B 37.5%
- 7
Llama 3.1-8B 34.5%
- 8 Granite-3.1-2B-Instruct 33.3%
- 9
Gemma 2 2B IT 32%
- 10 Granite-3.1-3B-A800M-Base 29.9%
Hard
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 62.5%
- 2 gpt-oss-120b 62%
- 3
Gemma 2 27B 41%
- 4
Gemma 2 9B 35.7%
- 5
Qwen2.5-3B ↑1 31.6%
- 6 gpt-oss-20b ↓1 31.4%
- 7 Granite-3.1-2B-Instruct ↑1 30.8%
- 8
Gemma 2 2B IT ↑1 30.4%
- 9
Llama 3.1-8B ↓2 29%
- 10
Llama 3.2 1B ↑2 27.7%
Assertion/Reason
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 71%
- 2 gpt-oss-120b 62.8%
- 3 Granite-3.1-2B-Instruct ↑5 48.9%
- 4
Gemma 3 270M ↑9 47.7%
- 5 Granite-3.1-3B-A800M-Base ↑5 44.4%
- 6
Gemma 2 2B IT ↑3 42%
- 7
Llama 3.2 1B ↑5 39.2%
- 8
Gemma 2 9B ↓4 38.9%
- 9
Gemma 2 27B ↓6 37%
- 10
Qwen2.5-3B ↓4 31.9%
Fill in the Blanks
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 59.2%
- 2
Gemma 2 27B ↑1 40.8%
- 3
Gemma 2 9B ↑1 40.8%
- 4 gpt-oss-120b ↓2 40.8%
- 5
Llama 3.1-8B ↑2 34.7%
- 6
Qwen2.5-3B 28.6%
- 7 Granite-3.1-3B-A800M-Base ↑3 28.6%
- 8
Gemma 2 2B IT ↑1 26.5%
- 9 gpt-oss-20b ↓4 26.5%
- 10
Gemma 3 270M Base ↑4 26.5%
Multiple Choice
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 73.1%
- 2 gpt-oss-120b 70.1%
- 3
Gemma 2 27B 61%
- 4
Gemma 2 9B 56%
- 5 gpt-oss-20b 46.7%
- 6
Qwen2.5-3B 44.1%
- 7
Llama 3.1-8B 39.5%
- 8 Granite-3.1-2B-Instruct 35.9%
- 9
Gemma 2 2B IT 34.6%
- 10
Llama 3.2 3B ↑1 32.7%
Match the Column
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 67.8%
- 2 gpt-oss-120b 64.2%
- 3
Gemma 2 27B 35.2%
- 4 gpt-oss-20b ↑1 29.6%
- 5
Gemma 2 2B IT ↑4 29%
- 6
Llama 3.2 1B ↑6 28.7%
- 7 Granite-3.1-2B-Instruct ↑1 28.7%
- 8
Gemma 2 9B ↓4 28.4%
- 9
Qwen2.5-3B ↓3 28.1%
- 10
Llama 3.2 3B ↑1 26.4%
Rearrange the Sequence
14 models we track have a score here.
- 1
Qwen3-235B-A22B-Instruct Jul 2025 75.1%
- 2 gpt-oss-120b 72.7%
- 3
Gemma 2 27B 46.9%
- 4
Qwen2.5-3B ↑2 37.3%
- 5 gpt-oss-20b 35.4%
- 6
Gemma 2 9B ↓2 34.9%
- 7 Granite-3.1-2B-Instruct ↑1 33.5%
- 8
Gemma 2 2B IT ↑1 29.7%
- 9
Llama 3.1-8B ↓2 28.2%
- 10
Llama 3.2 3B ↑1 27.8%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The BhashaBench-Krishi 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 BhashaBench-Krishi 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.
Versions
More in Knowledge
- MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects.
Intern-S2-397B 89.8%
- MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy.
Seed1.8 92.3%
- SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy.
Qwen 3.6 Max (Preview) 73.9%
- WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1.
Tencent Hy4 preview 83.9%
- BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly.
Gemma 4 31B IT 74.4%
- HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%
Data audit
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (31 entries).
- Pass Models scored 14 models we track, plus 17 other entries on its leaderboard.
- Note Newest score The newest came 15 months ago, in August 2025, for Gemma 3 270M. That’s expected for a test replaced by BhashaBench-Multi (Krishi).
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is gpt-oss-120b (#126).
- Pass Headroom The best is 72.7%, 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 and 17 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.