Benchmark Almanac

SciPredict

Gemini 3 Pro 25.3% leads since January 2026
Not in model cards

Models predict the outcomes of 405 real physics, biology and chemistry experiments published after March 31, 2025, with and without expert background knowledge.

Also measures calibration. The repo paper link was still a placeholder on November 1st, 2026.

Made by
Scale AI
Released
January 2026
Kind
Research science
Level
Expert
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Strongest model tested
GPT-5.2 #50 on the capability index
Last model added
December 2025
Scores refreshed
November 12th, 2026
15
Entries on its own leaderboard, 12 models we track

Gemini 3 Pro leads with 25.3% among models released by December 2025

Best score per entry from the benchmark's own leaderboard, read on November 12th, 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.

The newest model with a score here came out in December 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

15 entries. Scroll sideways for the rest.

  1. 25.3 Gemini 3 Pro
  2. 23.1 Claude Opus 4.5
  3. 22.6 Claude Sonnet 4.5
  4. 22.2 Claude Opus 4.1
  5. 22.2 Gemini 3 Flash
  6. 20.6 GPT-5.2
  7. 19.8 o3-mini
  8. 19.2 DeepSeek-V3 (snapshot unspecified)
  9. 18.2 Llama 3.3 70B
  10. 17.9 o3
  11. 17.0 gemini-2.5-pro (snapshot unspecified)
  12. 17.0 Qwen3-32B
  13. 16.6 Qwen3-235B
  14. 16.2 o4-mini
  15. 14.7 Llama 3.1-8B
  1. 1 Gemini 3 Pro 25.3%
  2. 2 Claude Opus 4.5 23.1%
  3. 3 Claude Sonnet 4.5 22.6%
  4. 4 Claude Opus 4.1 22.2%
  5. 5 Gemini 3 Flash 22.2%
  6. 6 GPT-5.2 20.6%
  7. 7 o3-mini 19.8%
  8. 8 DeepSeek-V3 (snapshot unspecified) 19.2%
  9. 9 Llama 3.3 70B 18.2%
  10. 10 o3 17.9%
  11. 11 gemini-2.5-pro (snapshot unspecified) 17%
  12. 12 Qwen3-32B 17%
  13. 13 Qwen3-235B 16.6%
  14. 14 o4-mini 16.2%
  15. 15 Llama 3.1-8B 14.7%

Source: Official SciPredict leaderboard

#EntryScoreSettingDateSource
1 Gemini 3 Pro 25.3% Nov 2025 Official leaderboard
2 Claude Opus 4.5 23.1% Nov 2025 Official leaderboard
3 Claude Sonnet 4.5 22.6% Sep 2025 Official leaderboard
4 Claude Opus 4.1 22.2% Aug 2025 Official leaderboard
5 Gemini 3 Flash 22.2% Dec 2025 Official leaderboard
6 GPT-5.2 20.6% Dec 2025 Official leaderboard
7 o3-mini 19.8% Jan 2025 Official leaderboard
8 DeepSeek-V3 (snapshot unspecified) 19.2% Official leaderboard
9 Llama 3.3 70B 18.2% Dec 2024 Official leaderboard
10 o3 17.9% Apr 2025 Official leaderboard
11 gemini-2.5-pro (snapshot unspecified) 17% Official leaderboard
12 Qwen3-32B 17% Apr 2025 Official leaderboard
13 Qwen3-235B 16.6% Official leaderboard
14 o4-mini 16.2% Apr 2025 Official leaderboard
15 Llama 3.1-8B 14.7% Jul 2024 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 Not tested since Jan 2026
Gemini 3 Pro 25.3%

Source: Official SciPredict leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3 Pro Claude Opus 4.1 o3-mini Llama 3.1-8B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemini 3 Pro Llama 3.3 70B

Source: Official SciPredict leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

More in Science and reasoning

All 92 in Science and reasoning
  • GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics. Gemini 4 Carbon 96.5%
  • Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
  • FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics. Intern-S2-397B 87%
  • BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Fable 5.2 83.7%
  • HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels. Gemini 3 Pro 78.3%
  • LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%