Benchmark Almanac

LLM Agronomy Benchmark Image Identification (four choices)

Gemini 3 Flash 90% leads since June 2026
Not in model cards

Models choose among four labels for each photo of a crop disease, insect pest, or weed. One accuracy score combines the 150 questions.

The June 16 leaderboard reports image accuracy for 39 model configurations. The overall text score excludes these questions.

Made by
gbstox
Released
June 2026
Kind
General visual questions
Level
University
Scored
Checked automatically
Answers
Multiple choice
Tasks
Kept private
Measures
Capability
Language
English
In and out
Text, Images → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
A host
In views
Vision
Strongest model tested
GPT-5.4 Pro #19 on the capability index
Last model added
June 2026
Scores refreshed
November 20th, 2026
38
Entries on its own leaderboard, 37 models we track

Gemini 3 Flash leads with 90% among models released by June 2026 Gemini 3 Flash leads with 90% among models released by June 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 June 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 Flash 90%
  2. 2 Gemini 3.1 Pro 88%
  3. 3 Gemini 2.5 Pro Jun 2025 87.3%
  4. 4 o3-pro 87.3%
  5. 5 Gemini 3.1 Flash Image 87.3%
  6. 6 o3 84.7%
  7. 7 Kimi K2.5 83.3%
  8. 8 GPT-5.2 82.7%
  9. 9 GPT-5.2 Pro 82.7%
  10. 10 Gemini 3.1 Flash-Lite 82.7%
  11. 11 GPT-5.5 82%
  12. 12 Qwen3.5-Plus 81.3%
  13. 13 GPT-5.4 Pro 81.3%
  14. 14 Qwen3.7-Plus 81.3%
  15. 15 Qwen3.5 397B-A17B 80.7%
  16. 16 GPT-5 mini 80%
  17. 17 GPT-5.3 Codex 80%
  18. 18 bytedance-seed/seed-1.6-flash 80%
  19. 19 Seed2.0 Mini 79.3%
  20. 20 Claude Opus 4.5 78.7%
  21. 21 Seed2.0 Lite (0215) 78.7%
  22. 22 Claude Opus 4.8 78.7%
  23. 23 Doubao-Seed-1.6 78%
  24. 24 GPT-5.2 Codex 77.3%
  25. 25 GPT-5.3 Instant 73.3%
  26. 26 Qwen3.5-Flash 72.7%
  27. 27 GPT-5.4 72.7%
  28. 28 Claude Opus 4.1 68%
  29. 29 Claude Opus 4 67.3%
  30. 30 Gemma 3 27B 66%
  31. 31 Qwen3.5-27B 65.3%
  32. 32 Llama 4 Maverick 64.7%
  33. 33 Claude Sonnet 4 64%
  34. 34 MiniMax-M3 62%
  35. 35 GPT-4o Mar 2025 61.3%
  36. 36 GPT-4o mini 60.7%
  37. 37 Llama 4 Scout 52%
  38. 38 Gemma 3 4B 44.7%

Source: Official LLM Agronomy Benchmark Image Identification (four choices) leaderboard

Source: Official LLM Agronomy Benchmark Image Identification (four choices) leaderboard

#EntryScoreSettingReleasedRun onSource
1 Gemini 3 Flash 90% Dec 2025 Jun 2026 Official leaderboard
2 Gemini 3.1 Pro 88% Feb 2026 Jun 2026 Official leaderboard
3 Gemini 2.5 Pro Jun 2025 87.3% Jun 2025 Jun 2026 Official leaderboard
4 o3-pro 87.3% Jun 2025 Jun 2026 Official leaderboard
5 Gemini 3.1 Flash Image 87.3% Feb 2026 Jun 2026 Official leaderboard
6 o3 84.7% Apr 2025 Jun 2026 Official leaderboard
7 Kimi K2.5 83.3% Jan 2026 Jun 2026 Official leaderboard
8 GPT-5.2 82.7% Dec 2025 Jun 2026 Official leaderboard
9 GPT-5.2 Pro 82.7% Dec 2025 Jun 2026 Official leaderboard
10 Gemini 3.1 Flash-Lite 82.7% Mar 2026 Jun 2026 Official leaderboard
11 GPT-5.5 82% Apr 2026 Jun 2026 Official leaderboard
12 Qwen3.5-Plus 81.3% Feb 2026 Jun 2026 Official leaderboard
13 GPT-5.4 Pro 81.3% Mar 2026 Jun 2026 Official leaderboard
14 Qwen3.7-Plus 81.3% Jun 2026 Jun 2026 Official leaderboard
15 Qwen3.5 397B-A17B 80.7% Feb 2026 Jun 2026 Official leaderboard
16 GPT-5 mini 80% Aug 2025 Jun 2026 Official leaderboard
17 GPT-5.3 Codex 80% Feb 2026 Jun 2026 Official leaderboard
18 bytedance-seed/seed-1.6-flash closed weights 80% Jun 2026 Official leaderboard
19 Seed2.0 Mini 79.3% Feb 2026 Jun 2026 Official leaderboard
20 Claude Opus 4.5 78.7% Nov 2025 Jun 2026 Official leaderboard
21 Seed2.0 Lite (0215) 78.7% Feb 2026 Jun 2026 Official leaderboard
22 Claude Opus 4.8 78.7% May 2026 Jun 2026 Official leaderboard
23 Doubao-Seed-1.6 78% Jun 2025 Jun 2026 Official leaderboard
24 GPT-5.2 Codex 77.3% Dec 2025 Jun 2026 Official leaderboard
25 GPT-5.3 Instant 73.3% Mar 2026 Jun 2026 Official leaderboard
26 Qwen3.5-Flash 72.7% Feb 2026 Jun 2026 Official leaderboard
27 GPT-5.4 72.7% Mar 2026 Jun 2026 Official leaderboard
28 Claude Opus 4.1 68% Aug 2025 Jun 2026 Official leaderboard
29 Claude Opus 4 67.3% May 2025 Jun 2026 Official leaderboard
30 Gemma 3 27B 66% Mar 2025 Jun 2026 Official leaderboard
31 Qwen3.5-27B 65.3% Feb 2026 Jun 2026 Official leaderboard
32 Llama 4 Maverick 64.7% Apr 2025 Jun 2026 Official leaderboard
33 Claude Sonnet 4 64% May 2025 Jun 2026 Official leaderboard
34 MiniMax-M3 62% Jun 2026 Jun 2026 Official leaderboard
35 GPT-4o Mar 2025 61.3% Mar 2025 Jun 2026 Official leaderboard
36 GPT-4o mini 60.7% Jul 2024 Jun 2026 Official leaderboard
37 Llama 4 Scout 52% Apr 2025 Jun 2026 Official leaderboard
38 Gemma 3 4B 44.7% Mar 2025 Jun 2026 Official leaderboard
#ModelScoreSettingReleasedRun onSource
1 Gemini 3 Flash 90% Dec 2025 Jun 2026 Official leaderboard
2 Gemini 3.1 Pro 88% Feb 2026 Jun 2026 Official leaderboard
3 Gemini 2.5 Pro Jun 2025 87.3% Jun 2025 Jun 2026 Official leaderboard
4 o3-pro 87.3% Jun 2025 Jun 2026 Official leaderboard
5 Gemini 3.1 Flash Image 87.3% Feb 2026 Jun 2026 Official leaderboard
6 o3 84.7% Apr 2025 Jun 2026 Official leaderboard
7 Kimi K2.5 83.3% Jan 2026 Jun 2026 Official leaderboard
8 GPT-5.2 82.7% Dec 2025 Jun 2026 Official leaderboard
9 GPT-5.2 Pro 82.7% Dec 2025 Jun 2026 Official leaderboard
10 Gemini 3.1 Flash-Lite 82.7% Mar 2026 Jun 2026 Official leaderboard
11 GPT-5.5 82% Apr 2026 Jun 2026 Official leaderboard
12 Qwen3.5-Plus 81.3% Feb 2026 Jun 2026 Official leaderboard
13 GPT-5.4 Pro 81.3% Mar 2026 Jun 2026 Official leaderboard
14 Qwen3.7-Plus 81.3% Jun 2026 Jun 2026 Official leaderboard
15 Qwen3.5 397B-A17B 80.7% Feb 2026 Jun 2026 Official leaderboard
16 GPT-5 mini 80% Aug 2025 Jun 2026 Official leaderboard
17 GPT-5.3 Codex 80% Feb 2026 Jun 2026 Official leaderboard
18 Seed2.0 Mini 79.3% Feb 2026 Jun 2026 Official leaderboard
19 Claude Opus 4.5 78.7% Nov 2025 Jun 2026 Official leaderboard
20 Seed2.0 Lite (0215) 78.7% Feb 2026 Jun 2026 Official leaderboard
21 Claude Opus 4.8 78.7% May 2026 Jun 2026 Official leaderboard
22 Doubao-Seed-1.6 78% Jun 2025 Jun 2026 Official leaderboard
23 GPT-5.2 Codex 77.3% Dec 2025 Jun 2026 Official leaderboard
24 GPT-5.3 Instant 73.3% Mar 2026 Jun 2026 Official leaderboard
25 Qwen3.5-Flash 72.7% Feb 2026 Jun 2026 Official leaderboard
26 GPT-5.4 72.7% Mar 2026 Jun 2026 Official leaderboard
27 Claude Opus 4.1 68% Aug 2025 Jun 2026 Official leaderboard
28 Claude Opus 4 67.3% May 2025 Jun 2026 Official leaderboard
29 Gemma 3 27B 66% Mar 2025 Jun 2026 Official leaderboard
30 Qwen3.5-27B 65.3% Feb 2026 Jun 2026 Official leaderboard
31 Llama 4 Maverick 64.7% Apr 2025 Jun 2026 Official leaderboard
32 Claude Sonnet 4 64% May 2025 Jun 2026 Official leaderboard
33 MiniMax-M3 62% Jun 2026 Jun 2026 Official leaderboard
34 GPT-4o Mar 2025 61.3% Mar 2025 Jun 2026 Official leaderboard
35 GPT-4o mini 60.7% Jul 2024 Jun 2026 Official leaderboard
36 Llama 4 Scout 52% Apr 2025 Jun 2026 Official leaderboard
37 Gemma 3 4B 44.7% Mar 2025 Jun 2026 Official leaderboard

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 Passed 85% in Jun 2026, by Gemini 3 Flash

Source: Official LLM Agronomy Benchmark Image Identification (four choices) leaderboard

Gemini 2.5 Pro on the official leaderboard passed 85% in June 2025

Models released before it came out already scored that high. 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% 2025 2026 Near the ceiling (85%) Gemini 3 Flash Gemini 2.5 Pro Gemma 3 27B GPT-4o mini 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3 Flash Gemma 3 27B

Source: Official LLM Agronomy Benchmark Image Identification (four choices) 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 Vision

All 636 in Vision
  • MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy. Qwen 3.6 Plus 86%
  • MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
  • CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
  • MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy. Seed2.1 Pro 90.7%
  • MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy. Qwen 3.8 Max 97.7%
  • MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy. Qwen 3.8 Max 85.9%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (38 entries).
  • Pass Models scored 37 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from June 2026, for Qwen3.7-Plus.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5.4 Pro (#19).
  • Note Headroom Saturated. The best is 90%, past 85% since June 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 38 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