Benchmark Almanac

Polaris-Bench

GPT-5.2 39.2% leads since May 2026
Not in model cards

Models solve the same 53 visual reasoning tasks in Cartesian and polar layouts. The benchmark compares accuracy across the paired layouts.

The authors publish leaderboard results for 14 multimodal models, including Gemini, GPT, Claude, and open models.

Made by
Google DeepMind
Released
May 2026
Kind
Spatial reasoning and counting
Level
School
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Language
English
In and out
Images, Text → Text
Interaction
Single turn
Items from
Generated
Refresh
One fixed set
Run by
Anyone
In views
Vision
Strongest model tested
Gemini 3.1 Pro #40 on the capability index
Last model added
April 2026
Scores refreshed
November 21st, 2026
14
Entries on its own leaderboard, 13 models we track

GPT-5.2 leads with 39.2% among models released by April 2026 GPT-5.2 leads with 39.2% among models released by April 2026

Best score per entry from the benchmark's own leaderboard, read on November 21st, 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 April 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 GPT-5.2 39.2%
  2. 2 Gemini 3.1 Pro 35.9%
  3. 3 Qwen3.5 397B-A17B 35%
  4. 4 Gemini 3 Flash 33.8%
  5. 5 Kimi K2.5 31.1%
  6. 6 Gemma 4 31B IT 31%
  7. 7 Claude Sonnet 4.6 25.9%
  8. 8 Gemini 2.5 Pro Jun 2025 25.3%
  9. 9 Gemini 3.1 Flash-Lite 24.6%
  10. 10 Gemma 4 26B A4B 22.9%
  11. 11 Grok 4 Fast reasoning 22.3%
  12. 12 Grok 4 21.8%
  13. 13 Gemini 2.5 Flash Sep 2025 21.1%
  14. 14 Mistral-Small-2503 19%

Source: Official Polaris-Bench leaderboard

#EntryScoreSettingDateSource
1 GPT-5.2 39.2% Dec 2025 Official leaderboard
2 Gemini 3.1 Pro 35.9% Feb 2026 Official leaderboard
3 Qwen3.5 397B-A17B 35% Feb 2026 Official leaderboard
4 Gemini 3 Flash 33.8% Dec 2025 Official leaderboard
5 Kimi K2.5 31.1% Jan 2026 Official leaderboard
6 Gemma 4 31B IT 31% Apr 2026 Official leaderboard
7 Claude Sonnet 4.6 25.9% Feb 2026 Official leaderboard
8 Gemini 2.5 Pro Jun 2025 25.3% Jun 2025 Official leaderboard
9 Gemini 3.1 Flash-Lite 24.6% Mar 2026 Official leaderboard
10 Gemma 4 26B A4B 22.9% Apr 2026 Official leaderboard
11 Grok 4 Fast 22.3% reasoning Sep 2025 Official leaderboard
12 Grok 4 21.8% Jul 2025 Official leaderboard
13 Gemini 2.5 Flash Sep 2025 21.1% Sep 2025 Official leaderboard
14 Mistral-Small-2503 open weights 19% Official leaderboard
#ModelScoreSettingDateSource
1 GPT-5.2 39.2% Dec 2025 Official leaderboard
2 Gemini 3.1 Pro 35.9% Feb 2026 Official leaderboard
3 Qwen3.5 397B-A17B 35% Feb 2026 Official leaderboard
4 Gemini 3 Flash 33.8% Dec 2025 Official leaderboard
5 Kimi K2.5 31.1% Jan 2026 Official leaderboard
6 Gemma 4 31B IT 31% Apr 2026 Official leaderboard
7 Claude Sonnet 4.6 25.9% Feb 2026 Official leaderboard
8 Gemini 2.5 Pro Jun 2025 25.3% Jun 2025 Official leaderboard
9 Gemini 3.1 Flash-Lite 24.6% Mar 2026 Official leaderboard
10 Gemma 4 26B A4B 22.9% Apr 2026 Official leaderboard
11 Grok 4 Fast 22.3% reasoning Sep 2025 Official leaderboard
12 Grok 4 21.8% Jul 2025 Official leaderboard
13 Gemini 2.5 Flash Sep 2025 21.1% Sep 2025 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: Cartesian and Polar. Each ranks models on one part of the test. Gemini 3.1 Pro leads Cartesian and GPT-5.2 leads Polar. Arrows show how far a model moved from its overall rank among the models we track.

Cartesian

Accuracy on the Cartesian layout. 13 models we track have a score here.

  1. 1 Gemini 3.1 Pro ↑1 82.6%
  2. 2 GPT-5.2 ↓1 77.4%
  3. 3 Qwen3.5 397B-A17B 72.9%
  4. 4 Gemini 3 Flash 71%
  5. 5 Kimi K2.5 69%
  6. 6 Gemma 4 31B IT 60.5%
  7. 7 Gemma 4 26B A4B ↑3 47.2%
  8. 8 Gemini 3.1 Flash-Lite ↑1 46.8%
  9. 9 Claude Sonnet 4.6 ↓2 44.4%
  10. 10 Gemini 2.5 Pro Jun 2025 ↓2 38.4%

Number one over time

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

OpenAI
Own leaderboard Not tested since May 2026
GPT-5.2 39.2%

Source: Official Polaris-Bench 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% 2025 2026 Near the ceiling (85%) GPT-5.2 Gemini 2.5 Pro 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) GPT-5.2 Gemini 2.5 Pro

Source: Official Polaris-Bench 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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 7 months ago, in April 2026, for Gemma 4 26B A4B. 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 (14 entries).
  • Pass Models scored 13 models we track, plus 1 other entry on its leaderboard.
  • Warning Newest score The newest came 7 months ago, in April 2026, for Gemma 4 26B A4B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3.1 Pro (#40).
  • Pass Headroom The best is 39.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on google-deepmind.github.io: 13 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 21st, 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