Benchmark Almanac

OpenLMM Spatial Leaderboard

GPT-5 62.5 leads since August 2025
Not in model cards

Models answer spatial reasoning tasks involving images. OpenCompass compares performance across the component evaluations.

The leaderboard aggregates results from component evaluations.

Made by
OpenCompass
Released
June 2025
Kind
Composite indices
Level
University
Scored
Settled by events
Answers
Short answer
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Images → Text
Interaction
Single turn
Items from
Another benchmark
Refresh
One fixed set
Run by
Anyone
In views
Vision
Strongest model tested
GPT-5 #71 on the capability index
Last model added
August 2025
Scores refreshed
November 20th, 2026
59
Entries on its own leaderboard, 51 models we track

GPT-5 leads with 62.5 among models released by August 2025 GPT-5 leads with 62.5 among models released by August 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 August 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 GPT-5 Standard ✓ 62.5
  2. 2 Seed1.5-VL Standard ✓ 57.3
  3. 3 GPT-5 mini Standard ✓ 57.1
  4. 4 GPT-4.1 Standard ✓ 55.2
  5. 5 ChatGPT-4o-latest Standard ✓ 54.6
  6. 6 Gemini 2.5 Flash Jun 2025, Standard ✓ 54.3
  7. 7 GPT-4o (1120, detail-high) Standard ✓ 53.9
  8. 8 Qwen2-VL-72B-Instruct Standard ✓ 53.1
  9. 9 InternVL3-38B Standard ✓ 52.8
  10. 10 InternVL3-78B Standard ✓ 52.8
  11. 11 Step-1o Standard ✓ 52.0
  12. 12 Gemini 2.0 Flash Dec 2024, Standard ✓ 51.8
  13. 13 Qwen2.5-VL-72B Standard ✓ 51.4
  14. 14 GPT-4.1 mini Standard ✓ 51.1
  15. 15 Qwen2.5-VL-32B Standard ✓ 51.0
  16. 16 Gemini 1.5 Pro Sep 2024, Standard ✓ 50.8
  17. 17 MiMo-VL-7B-RL Standard ✓ 50.5
  18. 18 Seed1.6-Flash Standard ✓ 49.3
  19. 19 InternVL2_5-78B Standard ✓ 49.2
  20. 20 Claude 3.7 Sonnet Standard ✓ 49.1
  21. 21 InternVL3-14B Standard ✓ 49.1
  22. 22 Claude 3.5 Sonnet (October 2024) Standard ✓ 49.0
  23. 23 Ovis2-34B Standard ✓ 48.9
  24. 24 GPT-5 nano Standard ✓ 48.2
  25. 25 Gemini 1.5 Flash Sep 2024, Standard ✓ 47.6
  26. 26 Ovis2-16B Standard ✓ 47.5
  27. 27 GLM-4v-Plus-20250111 Standard ✓ 47.4
  28. 28 LLaVA-OneVision 72B Standard ✓ 47.2
  29. 29 InternVL3-8B Standard ✓ 46.7
  30. 30 Qwen2-VL-7B-Instruct Standard ✓ 45.9
  31. 31 Gemini 2.5 Pro Mar 2025, Circular ✓ 45.0
  32. 32 Claude4-Sonnet Standard ✓ 44.4
  33. 33 Gemma 3 27B Standard ✓ 44.1
  34. 34 Kimi-VL-A3B-Instruct Standard ✓ 44.0
  35. 35 SAIL-VL-1.6-8B Standard ✓ 43.7
  36. 36 Ovis2-8B Standard ✓ 43.5
  37. 37 InternVL2.5-8B Standard ✓ 43.3
  38. 38 Qwen2.5-VL-7B Standard ✓ 43.3
  39. 39 DeepSeek-VL2 Standard ✓ 42.7
  40. 40 Gemma 3 12B Standard ✓ 42.3
  41. 41 Phi-4-multimodal Standard ✓ 42.1
  42. 42 GPT-4o-mini (0718, detail-high) Standard ✓ 42.0
  43. 43 InternLM-XComposer2.5-7B Standard ✓ 41.7
  44. 44 MiniCPM-o-2_6 Standard ✓ 41.5
  45. 45 DeepSeek-VL2-Small Standard ✓ 40.9
  46. 46 LLaVA-OneVision 7B Standard ✓ 40.8
  47. 47 SAIL-VL-1.5-2B Standard ✓ 39.8
  48. 48 Pixtral 12B Standard ✓ 39.5
  49. 49 GPT-4.1 nano Standard ✓ 39.0
  50. 50 Qwen2.5-VL-3B Standard ✓ 38.5
  51. 51 InternVL3-2B Standard ✓ 38.4
  52. 52 Phi-3.5-vision-instruct Standard ✓ 38.2
  53. 53 InternVL3-1B Standard ✓ 38.1
  54. 54 Gemma 3 4B Standard ✓ 37.4
  55. 55 Idefics3-8B-Llama3 Standard ✓ 33.1
  56. 56 Janus-Pro-7B Standard ✓ 32.4
  57. 57 SmolVLM2-2.2B-Instruct Standard ✓ 30.1
  58. 58 llava-v1.6-vicuna-7b Standard ✓ 19.9
  59. 59 llava-v1.6-vicuna-13b Standard ✓ 19.6

Source: Official OpenLMM Spatial Leaderboard leaderboard

  1. 1 GPT-5 Standard ✓ 62.5
  2. 2 Seed1.5-VL Standard ✓ 57.3
  3. 3 GPT-5 mini Standard ✓ 57.1
  4. 4 GPT-4.1 Standard ✓ 55.2
  5. 5 Gemini 2.5 Flash Jun 2025, Standard ✓ 54.3
  6. 6 Qwen2-VL-72B-Instruct Standard ✓ 53.1
  7. 7 InternVL3-38B Standard ✓ 52.8
  8. 8 InternVL3-78B Standard ✓ 52.8
  9. 9 Gemini 2.0 Flash Dec 2024, Standard ✓ 51.8
  10. 10 Qwen2.5-VL-72B Standard ✓ 51.4
  11. 11 GPT-4.1 mini Standard ✓ 51.1
  12. 12 Qwen2.5-VL-32B Standard ✓ 51.0
  13. 13 Gemini 1.5 Pro Sep 2024, Standard ✓ 50.8
  14. 14 MiMo-VL-7B-RL Standard ✓ 50.5
  15. 15 InternVL2_5-78B Standard ✓ 49.2
  16. 16 Claude 3.7 Sonnet Standard ✓ 49.1
  17. 17 InternVL3-14B Standard ✓ 49.1
  18. 18 Claude 3.5 Sonnet (October 2024) Standard ✓ 49.0
  19. 19 Ovis2-34B Standard ✓ 48.9
  20. 20 GPT-5 nano Standard ✓ 48.2
  21. 21 Gemini 1.5 Flash Sep 2024, Standard ✓ 47.6
  22. 22 Ovis2-16B Standard ✓ 47.5
  23. 23 LLaVA-OneVision 72B Standard ✓ 47.2
  24. 24 InternVL3-8B Standard ✓ 46.7
  25. 25 Qwen2-VL-7B-Instruct Standard ✓ 45.9
  26. 26 Gemini 2.5 Pro Mar 2025, Circular ✓ 45.0
  27. 27 Gemma 3 27B Standard ✓ 44.1
  28. 28 Kimi-VL-A3B-Instruct Standard ✓ 44.0
  29. 29 SAIL-VL-1.6-8B Standard ✓ 43.7
  30. 30 Ovis2-8B Standard ✓ 43.5
  31. 31 InternVL2.5-8B Standard ✓ 43.3
  32. 32 Qwen2.5-VL-7B Standard ✓ 43.3
  33. 33 DeepSeek-VL2 Standard ✓ 42.7
  34. 34 Gemma 3 12B Standard ✓ 42.3
  35. 35 Phi-4-multimodal Standard ✓ 42.1
  36. 36 InternLM-XComposer2.5-7B Standard ✓ 41.7
  37. 37 MiniCPM-o-2_6 Standard ✓ 41.5
  38. 38 DeepSeek-VL2-Small Standard ✓ 40.9
  39. 39 LLaVA-OneVision 7B Standard ✓ 40.8
  40. 40 Pixtral 12B Standard ✓ 39.5
  41. 41 GPT-4.1 nano Standard ✓ 39.0
  42. 42 Qwen2.5-VL-3B Standard ✓ 38.5
  43. 43 InternVL3-2B Standard ✓ 38.4
  44. 44 Phi-3.5-vision-instruct Standard ✓ 38.2
  45. 45 InternVL3-1B Standard ✓ 38.1
  46. 46 Gemma 3 4B Standard ✓ 37.4
  47. 47 Idefics3-8B-Llama3 Standard ✓ 33.1
  48. 48 Janus-Pro-7B Standard ✓ 32.4
  49. 49 SmolVLM2-2.2B-Instruct Standard ✓ 30.1
  50. 50 llava-v1.6-vicuna-7b Standard ✓ 19.9
  51. 51 llava-v1.6-vicuna-13b Standard ✓ 19.6

Source: Official OpenLMM Spatial Leaderboard leaderboard

#EntryScoreSettingReleasedRun onLinkSource
1 GPT-5 ✓ 62.5 Standard Aug 2025 Aug 2025 openai.com Official leaderboard
2 Seed1.5-VL ✓ 57.3 Standard May 2025 Jun 2025 github.com Official leaderboard
3 GPT-5 mini ✓ 57.1 Standard Aug 2025 Aug 2025 openai.com Official leaderboard
4 GPT-4.1 ✓ 55.2 Standard Apr 2025 Apr 2025 openai.com Official leaderboard
5 ChatGPT-4o-latest ✓ closed weights 54.6 Standard Apr 2025 openai.com Official leaderboard
6 Gemini 2.5 Flash Jun 2025 ✓ 54.3 Standard Jun 2025 Jun 2025 deepmind.google Official leaderboard
7 GPT-4o (1120, detail-high) ✓ closed weights 53.9 Standard Dec 2024 openai.com Official leaderboard
8 Qwen2-VL-72B-Instruct ✓ 53.1 Standard Aug 2024 Oct 2024 huggingface.co Official leaderboard
9 InternVL3-38B ✓ 52.8 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
10 InternVL3-78B ✓ 52.8 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
11 Step-1o ✓ closed weights 52.0 Standard Jan 2025 platform.stepfun.com Official leaderboard
12 Gemini 2.0 Flash Dec 2024 ✓ 51.8 Standard Dec 2024 Jan 2025 deepmind.google Official leaderboard
13 Qwen2.5-VL-72B ✓ 51.4 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
14 GPT-4.1 mini ✓ 51.1 Standard Apr 2025 Apr 2025 openai.com Official leaderboard
15 Qwen2.5-VL-32B ✓ 51.0 Standard Mar 2025 Jun 2025 huggingface.co Official leaderboard
16 Gemini 1.5 Pro Sep 2024 ✓ 50.8 Standard Sep 2024 Dec 2024 deepmind.google Official leaderboard
17 MiMo-VL-7B-RL ✓ 50.5 Standard May 2025 Jun 2025 huggingface.co Official leaderboard
18 Seed1.6-Flash ✓ closed weights 49.3 Standard Jun 2025 seed.bytedance.com Official leaderboard
19 InternVL2_5-78B ✓ 49.2 Standard Dec 2024 Dec 2024 huggingface.co Official leaderboard
20 Claude 3.7 Sonnet ✓ 49.1 Standard Feb 2025 Feb 2025 docs.anthropic.com Official leaderboard
21 InternVL3-14B ✓ 49.1 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
22 Claude 3.5 Sonnet (October 2024) ✓ 49.0 Standard Oct 2024 Nov 2024 docs.anthropic.com Official leaderboard
23 Ovis2-34B ✓ 48.9 Standard Feb 2025 Feb 2025 huggingface.co Official leaderboard
24 GPT-5 nano ✓ 48.2 Standard Aug 2025 Aug 2025 openai.com Official leaderboard
25 Gemini 1.5 Flash Sep 2024 ✓ 47.6 Standard Aug 2024 Dec 2024 deepmind.google Official leaderboard
26 Ovis2-16B ✓ 47.5 Standard Feb 2025 Feb 2025 huggingface.co Official leaderboard
27 GLM-4v-Plus-20250111 ✓ closed weights 47.4 Standard Jan 2025 open.bigmodel.cn Official leaderboard
28 LLaVA-OneVision 72B ✓ 47.2 Standard Aug 2024 Sep 2024 huggingface.co Official leaderboard
29 InternVL3-8B ✓ 46.7 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
30 Qwen2-VL-7B-Instruct ✓ 45.9 Standard Aug 2024 Sep 2024 huggingface.co Official leaderboard
31 Gemini 2.5 Pro Mar 2025 ✓ 45.0 Circular Mar 2025 Apr 2025 deepmind.google Official leaderboard
32 Claude4-Sonnet ✓ closed weights 44.4 Standard Jun 2025 anthropic.com Official leaderboard
33 Gemma 3 27B ✓ 44.1 Standard Mar 2025 Mar 2025 huggingface.co Official leaderboard
34 Kimi-VL-A3B-Instruct ✓ 44.0 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
35 SAIL-VL-1.6-8B ✓ 43.7 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
36 Ovis2-8B ✓ 43.5 Standard Feb 2025 Feb 2025 huggingface.co Official leaderboard
37 InternVL2.5-8B ✓ 43.3 Standard Dec 2024 Dec 2024 huggingface.co Official leaderboard
38 Qwen2.5-VL-7B ✓ 43.3 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
39 DeepSeek-VL2 ✓ 42.7 Standard Dec 2024 Jan 2025 huggingface.co Official leaderboard
40 Gemma 3 12B ✓ 42.3 Standard Mar 2025 Mar 2025 huggingface.co Official leaderboard
41 Phi-4-multimodal ✓ 42.1 Standard Feb 2025 Mar 2025 huggingface.co Official leaderboard
42 GPT-4o-mini (0718, detail-high) ✓ closed weights 42.0 Standard Jul 2024 openai.com Official leaderboard
43 InternLM-XComposer2.5-7B ✓ 41.7 Standard Jul 2024 Jul 2024 huggingface.co Official leaderboard
44 MiniCPM-o-2_6 ✓ 41.5 Standard Jan 2025 Jan 2025 huggingface.co Official leaderboard
45 DeepSeek-VL2-Small ✓ 40.9 Standard Dec 2024 Jan 2025 huggingface.co Official leaderboard
46 LLaVA-OneVision 7B ✓ 40.8 Standard Aug 2024 Sep 2024 huggingface.co Official leaderboard
47 SAIL-VL-1.5-2B ✓ open weights, 2.5B 39.8 Standard Apr 2025 huggingface.co Official leaderboard
48 Pixtral 12B ✓ 39.5 Standard Sep 2024 Sep 2024 huggingface.co Official leaderboard
49 GPT-4.1 nano ✓ 39.0 Standard Apr 2025 Apr 2025 openai.com Official leaderboard
50 Qwen2.5-VL-3B ✓ 38.5 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
51 InternVL3-2B ✓ 38.4 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
52 Phi-3.5-vision-instruct ✓ 38.2 Standard Aug 2024 Aug 2024 huggingface.co Official leaderboard
53 InternVL3-1B ✓ 38.1 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
54 Gemma 3 4B ✓ 37.4 Standard Mar 2025 Mar 2025 huggingface.co Official leaderboard
55 Idefics3-8B-Llama3 ✓ 33.1 Standard Aug 2024 Aug 2024 huggingface.co Official leaderboard
56 Janus-Pro-7B ✓ 32.4 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
57 SmolVLM2-2.2B-Instruct ✓ 30.1 Standard Feb 2025 Mar 2025 huggingface.co Official leaderboard
58 llava-v1.6-vicuna-7b ✓ 19.9 Standard Jan 2024 Mar 2024 llava-vl.github.io Official leaderboard
59 llava-v1.6-vicuna-13b ✓ 19.6 Standard Jan 2024 Mar 2024 llava-vl.github.io Official leaderboard
#ModelScoreSettingReleasedRun onLinkSource
1 GPT-5 ✓ 62.5 Standard Aug 2025 Aug 2025 openai.com Official leaderboard
2 Seed1.5-VL ✓ 57.3 Standard May 2025 Jun 2025 github.com Official leaderboard
3 GPT-5 mini ✓ 57.1 Standard Aug 2025 Aug 2025 openai.com Official leaderboard
4 GPT-4.1 ✓ 55.2 Standard Apr 2025 Apr 2025 openai.com Official leaderboard
5 Gemini 2.5 Flash Jun 2025 ✓ 54.3 Standard Jun 2025 Jun 2025 deepmind.google Official leaderboard
6 Qwen2-VL-72B-Instruct ✓ 53.1 Standard Aug 2024 Oct 2024 huggingface.co Official leaderboard
7 InternVL3-38B ✓ 52.8 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
8 InternVL3-78B ✓ 52.8 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
9 Gemini 2.0 Flash Dec 2024 ✓ 51.8 Standard Dec 2024 Jan 2025 deepmind.google Official leaderboard
10 Qwen2.5-VL-72B ✓ 51.4 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
11 GPT-4.1 mini ✓ 51.1 Standard Apr 2025 Apr 2025 openai.com Official leaderboard
12 Qwen2.5-VL-32B ✓ 51.0 Standard Mar 2025 Jun 2025 huggingface.co Official leaderboard
13 Gemini 1.5 Pro Sep 2024 ✓ 50.8 Standard Sep 2024 Dec 2024 deepmind.google Official leaderboard
14 MiMo-VL-7B-RL ✓ 50.5 Standard May 2025 Jun 2025 huggingface.co Official leaderboard
15 InternVL2_5-78B ✓ 49.2 Standard Dec 2024 Dec 2024 huggingface.co Official leaderboard
16 Claude 3.7 Sonnet ✓ 49.1 Standard Feb 2025 Feb 2025 docs.anthropic.com Official leaderboard
17 InternVL3-14B ✓ 49.1 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
18 Claude 3.5 Sonnet (October 2024) ✓ 49.0 Standard Oct 2024 Nov 2024 docs.anthropic.com Official leaderboard
19 Ovis2-34B ✓ 48.9 Standard Feb 2025 Feb 2025 huggingface.co Official leaderboard
20 GPT-5 nano ✓ 48.2 Standard Aug 2025 Aug 2025 openai.com Official leaderboard
21 Gemini 1.5 Flash Sep 2024 ✓ 47.6 Standard Aug 2024 Dec 2024 deepmind.google Official leaderboard
22 Ovis2-16B ✓ 47.5 Standard Feb 2025 Feb 2025 huggingface.co Official leaderboard
23 LLaVA-OneVision 72B ✓ 47.2 Standard Aug 2024 Sep 2024 huggingface.co Official leaderboard
24 InternVL3-8B ✓ 46.7 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
25 Qwen2-VL-7B-Instruct ✓ 45.9 Standard Aug 2024 Sep 2024 huggingface.co Official leaderboard
26 Gemini 2.5 Pro Mar 2025 ✓ 45.0 Circular Mar 2025 Apr 2025 deepmind.google Official leaderboard
27 Gemma 3 27B ✓ 44.1 Standard Mar 2025 Mar 2025 huggingface.co Official leaderboard
28 Kimi-VL-A3B-Instruct ✓ 44.0 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
29 SAIL-VL-1.6-8B ✓ 43.7 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
30 Ovis2-8B ✓ 43.5 Standard Feb 2025 Feb 2025 huggingface.co Official leaderboard
31 InternVL2.5-8B ✓ 43.3 Standard Dec 2024 Dec 2024 huggingface.co Official leaderboard
32 Qwen2.5-VL-7B ✓ 43.3 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
33 DeepSeek-VL2 ✓ 42.7 Standard Dec 2024 Jan 2025 huggingface.co Official leaderboard
34 Gemma 3 12B ✓ 42.3 Standard Mar 2025 Mar 2025 huggingface.co Official leaderboard
35 Phi-4-multimodal ✓ 42.1 Standard Feb 2025 Mar 2025 huggingface.co Official leaderboard
36 InternLM-XComposer2.5-7B ✓ 41.7 Standard Jul 2024 Jul 2024 huggingface.co Official leaderboard
37 MiniCPM-o-2_6 ✓ 41.5 Standard Jan 2025 Jan 2025 huggingface.co Official leaderboard
38 DeepSeek-VL2-Small ✓ 40.9 Standard Dec 2024 Jan 2025 huggingface.co Official leaderboard
39 LLaVA-OneVision 7B ✓ 40.8 Standard Aug 2024 Sep 2024 huggingface.co Official leaderboard
40 Pixtral 12B ✓ 39.5 Standard Sep 2024 Sep 2024 huggingface.co Official leaderboard
41 GPT-4.1 nano ✓ 39.0 Standard Apr 2025 Apr 2025 openai.com Official leaderboard
42 Qwen2.5-VL-3B ✓ 38.5 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
43 InternVL3-2B ✓ 38.4 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
44 Phi-3.5-vision-instruct ✓ 38.2 Standard Aug 2024 Aug 2024 huggingface.co Official leaderboard
45 InternVL3-1B ✓ 38.1 Standard Apr 2025 Apr 2025 huggingface.co Official leaderboard
46 Gemma 3 4B ✓ 37.4 Standard Mar 2025 Mar 2025 huggingface.co Official leaderboard
47 Idefics3-8B-Llama3 ✓ 33.1 Standard Aug 2024 Aug 2024 huggingface.co Official leaderboard
48 Janus-Pro-7B ✓ 32.4 Standard Jan 2025 Feb 2025 huggingface.co Official leaderboard
49 SmolVLM2-2.2B-Instruct ✓ 30.1 Standard Feb 2025 Mar 2025 huggingface.co Official leaderboard
50 llava-v1.6-vicuna-7b ✓ 19.9 Standard Jan 2024 Mar 2024 llava-vl.github.io Official leaderboard
51 llava-v1.6-vicuna-13b ✓ 19.6 Standard Jan 2024 Mar 2024 llava-vl.github.io Official leaderboard

Rankings by part

We read 60 parts from the benchmark’s own leaderboard in 5 groups: blink, lego, 3dsrbench, spatial457 and mmsibench. Each ranks models on one part of the test. GPT-5 leads 23 of them, and 15 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

BLINK

51 models we track have a score here.

  1. 1 Seed1.5-VL ↑1 72.8
  2. 2 GPT-5 ↓1 70.1
  3. 3 GPT-4.1 ↑1 68.0
  4. 4 GPT-5 mini ↓1 67.0
  5. 5 Gemini 2.5 Flash Jun 2025 66.6
  6. 6 InternVL3-78B ↑2 66.5
  7. 7 GPT-4.1 mini ↑4 64.6
  8. 8 InternVL3-38B ↓1 64.5
  9. 9 Gemini 2.0 Flash Dec 2024 63.9
  10. 10 Gemini 1.5 Pro Sep 2024 ↑3 63.8

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.

Lowest effort to highest, light to dark
GPT-5
Circular Standard
62.5
Seed1.5-VL
Circular Standard
57.3
GPT-5 mini
Circular Standard
57.1
GPT-4.1
Circular Standard
55.2
Gemini 2.5 Flash
Circular Standard
54.3
Qwen2-VL-72B-Instruct
Circular Standard
53.1
InternVL3-38B
Circular Standard
52.8
InternVL3-78B
Circular Standard
52.8
Gemini 2.0 Flash
Circular Standard
51.8
Qwen2.5-VL-72B
Circular Standard
51.4
GPT-4.1 mini
Circular Standard
51.1
Qwen2.5-VL-32B
Circular Standard
51.0

Source: Official OpenLMM Spatial Leaderboard leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-5 Standard 62.5
Seed1.5-VL Standard 57.3
GPT-5 mini Standard 57.1
GPT-4.1 Standard 55.2
Gemini 2.5 Flash Jun 2025 Standard 54.3
Qwen2-VL-72B-Instruct Standard 53.1
InternVL3-38B Standard 52.8
InternVL3-78B Standard 52.8
Gemini 2.0 Flash Dec 2024 Standard 51.8
Qwen2.5-VL-72B Standard 51.4
GPT-4.1 mini Standard 51.1
Qwen2.5-VL-32B Standard 51.0
Gemini 1.5 Pro Sep 2024 Standard 50.8
MiMo-VL-7B-RL Standard 50.5
InternVL2_5-78B Standard 49.2
Claude 3.7 Sonnet Standard 49.1
InternVL3-14B Standard 49.1
Claude 3.5 Sonnet (October 2024) Standard 49.0
Ovis2-34B Standard 48.9
GPT-5 nano Standard 48.2
Gemini 1.5 Flash Sep 2024 Standard 47.6
Ovis2-16B Standard 47.5
GPT-5 Circular 47.5
LLaVA-OneVision 72B Standard 47.2
InternVL3-8B Standard 46.7
Qwen2-VL-7B-Instruct Standard 45.9
Seed1.5-VL Circular 45.7
Gemini 2.5 Pro Mar 2025 Circular 45.0
Gemma 3 27B Standard 44.1
Kimi-VL-A3B-Instruct Standard 44.0
SAIL-VL-1.6-8B Standard 43.7
Ovis2-8B Standard 43.5
InternVL2.5-8B Standard 43.3
Qwen2.5-VL-7B Standard 43.3
DeepSeek-VL2 Standard 42.7
Gemma 3 12B Standard 42.3
Phi-4-multimodal Standard 42.1
InternLM-XComposer2.5-7B Standard 41.7
MiniCPM-o-2_6 Standard 41.5
InternVL3-78B Circular 41.3
InternVL3-38B Circular 41.2
DeepSeek-VL2-Small Standard 40.9
LLaVA-OneVision 7B Standard 40.8
GPT-4.1 Circular 40.5
GPT-5 mini Circular 40.0
Pixtral 12B Standard 39.5
Qwen2.5-VL-72B Circular 39.3
GPT-4.1 nano Standard 39.0
Qwen2.5-VL-3B Standard 38.5
InternVL3-2B Standard 38.4
Phi-3.5-vision-instruct Standard 38.2
InternVL3-1B Standard 38.1
InternVL2_5-78B Circular 37.6
Gemma 3 4B Standard 37.4
Gemini 2.5 Flash Jun 2025 Circular 36.9
Qwen2-VL-72B-Instruct Circular 35.9
Qwen2.5-VL-32B Circular 35.9
InternVL3-14B Circular 35.9
GPT-4.1 mini Circular 34.9
Gemini 2.0 Flash Dec 2024 Circular 34.8
Ovis2-16B Circular 34.6
Ovis2-34B Circular 34.0
LLaVA-OneVision 72B Circular 33.9
Claude 3.5 Sonnet (October 2024) Circular 33.4
Gemini 1.5 Pro Sep 2024 Circular 33.3
Idefics3-8B-Llama3 Standard 33.1
MiMo-VL-7B-RL Circular 33.0
Claude 3.7 Sonnet Circular 32.5
Janus-Pro-7B Standard 32.4
SAIL-VL-1.6-8B Circular 31.9
InternVL3-8B Circular 31.3
Ovis2-8B Circular 30.5
SmolVLM2-2.2B-Instruct Standard 30.1
MiniCPM-o-2_6 Circular 29.8
Kimi-VL-A3B-Instruct Circular 29.8
Qwen2.5-VL-7B Circular 29.7
InternVL2.5-8B Circular 29.6
InternLM-XComposer2.5-7B Circular 29.4
GPT-5 nano Circular 29.1
LLaVA-OneVision 7B Circular 28.3
DeepSeek-VL2 Circular 28.2
Phi-4-multimodal Circular 28.1
Gemini 1.5 Flash Sep 2024 Circular 27.2
DeepSeek-VL2-Small Circular 26.8
Gemma 3 27B Circular 26.7
Qwen2-VL-7B-Instruct Circular 25.8
Phi-3.5-vision-instruct Circular 24.7
InternVL3-2B Circular 23.7
Qwen2.5-VL-3B Circular 23.3
Gemma 3 12B Circular 23.1
Pixtral 12B Circular 21.7
InternVL3-1B Circular 20.1
Idefics3-8B-Llama3 Circular 20.0
llava-v1.6-vicuna-7b Standard 19.9
llava-v1.6-vicuna-13b Standard 19.6
SmolVLM2-2.2B-Instruct Circular 19.1
GPT-4.1 nano Circular 18.8
Gemma 3 4B Circular 15.7
Janus-Pro-7B Circular 15.2
llava-v1.6-vicuna-7b Circular 14.1
llava-v1.6-vicuna-13b Circular 12.7

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.

Lowest effort to highest, light to dark
GPT-5
Circular Standard
62.5
Seed1.5-VL
Circular Standard
57.3
GPT-5 mini
Circular Standard
57.1
GPT-4.1
Circular Standard
55.2
Gemini 2.5 Flash
Circular Standard
54.3
Qwen2-VL-72B-Instruct
Circular Standard
53.1
InternVL3-38B
Circular Standard
52.8
InternVL3-78B
Circular Standard
52.8
Gemini 2.0 Flash
Circular Standard
51.8
Qwen2.5-VL-72B
Circular Standard
51.4
GPT-4.1 mini
Circular Standard
51.1
Qwen2.5-VL-32B
Circular Standard
51.0

Source: Official OpenLMM Spatial Leaderboard leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-5 Standard 62.5
Seed1.5-VL Standard 57.3
GPT-5 mini Standard 57.1
GPT-4.1 Standard 55.2
Gemini 2.5 Flash Jun 2025 Standard 54.3
Qwen2-VL-72B-Instruct Standard 53.1
InternVL3-38B Standard 52.8
InternVL3-78B Standard 52.8
Gemini 2.0 Flash Dec 2024 Standard 51.8
Qwen2.5-VL-72B Standard 51.4
GPT-4.1 mini Standard 51.1
Qwen2.5-VL-32B Standard 51.0
Gemini 1.5 Pro Sep 2024 Standard 50.8
MiMo-VL-7B-RL Standard 50.5
InternVL2_5-78B Standard 49.2
Claude 3.7 Sonnet Standard 49.1
InternVL3-14B Standard 49.1
Claude 3.5 Sonnet (October 2024) Standard 49.0
Ovis2-34B Standard 48.9
GPT-5 nano Standard 48.2
Gemini 1.5 Flash Sep 2024 Standard 47.6
Ovis2-16B Standard 47.5
GPT-5 Circular 47.5
LLaVA-OneVision 72B Standard 47.2
InternVL3-8B Standard 46.7
Qwen2-VL-7B-Instruct Standard 45.9
Seed1.5-VL Circular 45.7
Gemini 2.5 Pro Mar 2025 Circular 45.0
Gemma 3 27B Standard 44.1
Kimi-VL-A3B-Instruct Standard 44.0
SAIL-VL-1.6-8B Standard 43.7
Ovis2-8B Standard 43.5
InternVL2.5-8B Standard 43.3
Qwen2.5-VL-7B Standard 43.3
DeepSeek-VL2 Standard 42.7
Gemma 3 12B Standard 42.3
Phi-4-multimodal Standard 42.1
InternLM-XComposer2.5-7B Standard 41.7
MiniCPM-o-2_6 Standard 41.5
InternVL3-78B Circular 41.3
InternVL3-38B Circular 41.2
DeepSeek-VL2-Small Standard 40.9
LLaVA-OneVision 7B Standard 40.8
GPT-4.1 Circular 40.5
GPT-5 mini Circular 40.0
Pixtral 12B Standard 39.5
Qwen2.5-VL-72B Circular 39.3
GPT-4.1 nano Standard 39.0
Qwen2.5-VL-3B Standard 38.5
InternVL3-2B Standard 38.4
Phi-3.5-vision-instruct Standard 38.2
InternVL3-1B Standard 38.1
InternVL2_5-78B Circular 37.6
Gemma 3 4B Standard 37.4
Gemini 2.5 Flash Jun 2025 Circular 36.9
Qwen2-VL-72B-Instruct Circular 35.9
Qwen2.5-VL-32B Circular 35.9
InternVL3-14B Circular 35.9
GPT-4.1 mini Circular 34.9
Gemini 2.0 Flash Dec 2024 Circular 34.8
Ovis2-16B Circular 34.6
Ovis2-34B Circular 34.0
LLaVA-OneVision 72B Circular 33.9
Claude 3.5 Sonnet (October 2024) Circular 33.4
Gemini 1.5 Pro Sep 2024 Circular 33.3
Idefics3-8B-Llama3 Standard 33.1
MiMo-VL-7B-RL Circular 33.0
Claude 3.7 Sonnet Circular 32.5
Janus-Pro-7B Standard 32.4
SAIL-VL-1.6-8B Circular 31.9
InternVL3-8B Circular 31.3
Ovis2-8B Circular 30.5
SmolVLM2-2.2B-Instruct Standard 30.1
MiniCPM-o-2_6 Circular 29.8
Kimi-VL-A3B-Instruct Circular 29.8
Qwen2.5-VL-7B Circular 29.7
InternVL2.5-8B Circular 29.6
InternLM-XComposer2.5-7B Circular 29.4
GPT-5 nano Circular 29.1
LLaVA-OneVision 7B Circular 28.3
DeepSeek-VL2 Circular 28.2
Phi-4-multimodal Circular 28.1
Gemini 1.5 Flash Sep 2024 Circular 27.2
DeepSeek-VL2-Small Circular 26.8
Gemma 3 27B Circular 26.7
Qwen2-VL-7B-Instruct Circular 25.8
Phi-3.5-vision-instruct Circular 24.7
InternVL3-2B Circular 23.7
Qwen2.5-VL-3B Circular 23.3
Gemma 3 12B Circular 23.1
Pixtral 12B Circular 21.7
InternVL3-1B Circular 20.1
Idefics3-8B-Llama3 Circular 20.0
llava-v1.6-vicuna-7b Standard 19.9
llava-v1.6-vicuna-13b Standard 19.6
SmolVLM2-2.2B-Instruct Circular 19.1
GPT-4.1 nano Circular 18.8
Gemma 3 4B Circular 15.7
Janus-Pro-7B Circular 15.2
llava-v1.6-vicuna-7b Circular 14.1
llava-v1.6-vicuna-13b Circular 12.7

2 models have held number one on the official leaderboard since its release in June 2025

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

ByteDance Seed OpenAI
Own leaderboard Not tested since Aug 2025
GPT-5 62.5

Source: Official OpenLMM Spatial 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
20.0 40.0 60.0 2024 2025 20.0 40.0 60.0 2024 2025

Source: Official OpenLMM Spatial 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

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 15 months ago, in August 2025, for GPT-5. 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 (59 entries).
  • Pass Models scored 51 models we track, plus 8 other entries on its leaderboard.
  • Warning Newest score The newest came 15 months ago, in August 2025, for GPT-5.
  • Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5 (#71).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on huggingface.co: 101 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