Benchmark Almanac

InteractScience

Claude Sonnet 4 41.5 leads since October 2025
Not in model cards

Models generate code for interactive scientific demonstrations. Evaluators test the program and assess its visual output.

Released
October 2025
Kind
Scientific code
Level
University
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Code
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5 #71 on the capability index
Last model added
September 2025
Scores refreshed
November 20th, 2026
30
Entries on its own leaderboard, 27 models we track

Claude Sonnet 4 leads with 41.5 among models released by September 2025 Claude Sonnet 4 leads with 41.5 among models released by September 2025

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. The authors report separate functional and visual measures. The headline is the overall functional test pass rate. 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 September 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 Claude Sonnet 4 41.5
  2. 2 Claude Opus 4 40.3
  3. 3 GPT-5 39.5
  4. 4 o4-mini 37.3
  5. 5 GPT-4.1 37.1
  6. 6 Gemini-2.5-Pro (snapshot unspecified) 35.3
  7. 7 o3 34.9
  8. 8 DeepSeek-R1 May 2025 33.9
  9. 9 Claude-3.5-Sonnet (snapshot unspecified) 33.3
  10. 10 Qwen3-235B-A22B-Instruct Jul 2025 33.3
  11. 11 Intern-S1 31.9
  12. 12 DeepSeek-V3 Mar 2025 31.7
  13. 13 Kimi K2 Sep 2025 31.6
  14. 14 Gemini 2.5 Flash Sep 2025 31.6
  15. 15 GLM-4.5 29.3
  16. 16 GPT-4o (snapshot unspecified) 28.3
  17. 17 gpt-oss-120b 28.0
  18. 18 Qwen2.5-Coder-32B-Instruct 27.2
  19. 19 Qwen3-32B 27.2
  20. 20 Qwen3-14B 24.1
  21. 21 Qwen2.5-VL-72B 23.7
  22. 22 Qwen2.5-Coder-14B-Instruct 22.5
  23. 23 Qwen3-8B 20.0
  24. 24 Llama 3.1-70B 18.7
  25. 25 gpt-oss-20b 15.2
  26. 26 Qwen3-4B 14.7
  27. 27 Qwen2.5-Coder (7B) 12.4
  28. 28 Llama 3.1-8B 11.3
  29. 29 Qwen2.5-VL-7B 7.47
  30. 30 Qwen3-1.7B 6.53

Source: Official InteractScience leaderboard

#EntryScoreSettingDateSource
1 Claude Sonnet 4 41.5 May 2025 Official leaderboard
2 Claude Opus 4 40.3 May 2025 Official leaderboard
3 GPT-5 39.5 Aug 2025 Official leaderboard
4 o4-mini 37.3 Apr 2025 Official leaderboard
5 GPT-4.1 37.1 Apr 2025 Official leaderboard
6 Gemini-2.5-Pro (snapshot unspecified) 35.3 Official leaderboard
7 o3 34.9 Apr 2025 Official leaderboard
8 DeepSeek-R1 May 2025 33.9 May 2025 Official leaderboard
9 Claude-3.5-Sonnet (snapshot unspecified) 33.3 Official leaderboard
10 Qwen3-235B-A22B-Instruct Jul 2025 33.3 Jul 2025 Official leaderboard
11 Intern-S1 31.9 Jul 2025 Official leaderboard
12 DeepSeek-V3 Mar 2025 31.7 Mar 2025 Official leaderboard
13 Kimi K2 Sep 2025 31.6 Sep 2025 Official leaderboard
14 Gemini 2.5 Flash Sep 2025 31.6 Sep 2025 Official leaderboard
15 GLM-4.5 29.3 Aug 2025 Official leaderboard
16 GPT-4o (snapshot unspecified) 28.3 Official leaderboard
17 gpt-oss-120b 28.0 Aug 2025 Official leaderboard
18 Qwen2.5-Coder-32B-Instruct 27.2 Nov 2024 Official leaderboard
19 Qwen3-32B 27.2 Apr 2025 Official leaderboard
20 Qwen3-14B 24.1 Apr 2025 Official leaderboard
21 Qwen2.5-VL-72B 23.7 Jan 2025 Official leaderboard
22 Qwen2.5-Coder-14B-Instruct 22.5 Nov 2024 Official leaderboard
23 Qwen3-8B 20.0 Apr 2025 Official leaderboard
24 Llama 3.1-70B 18.7 Jul 2024 Official leaderboard
25 gpt-oss-20b 15.2 Aug 2025 Official leaderboard
26 Qwen3-4B 14.7 Apr 2025 Official leaderboard
27 Qwen2.5-Coder (7B) 12.4 Sep 2024 Official leaderboard
28 Llama 3.1-8B 11.3 Jul 2024 Official leaderboard
29 Qwen2.5-VL-7B 7.47 Jan 2025 Official leaderboard
30 Qwen3-1.7B 6.53 Apr 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Sonnet 4 41.5 May 2025 Official leaderboard
2 Claude Opus 4 40.3 May 2025 Official leaderboard
3 GPT-5 39.5 Aug 2025 Official leaderboard
4 o4-mini 37.3 Apr 2025 Official leaderboard
5 GPT-4.1 37.1 Apr 2025 Official leaderboard
6 o3 34.9 Apr 2025 Official leaderboard
7 DeepSeek-R1 May 2025 33.9 May 2025 Official leaderboard
8 Qwen3-235B-A22B-Instruct Jul 2025 33.3 Jul 2025 Official leaderboard
9 Intern-S1 31.9 Jul 2025 Official leaderboard
10 DeepSeek-V3 Mar 2025 31.7 Mar 2025 Official leaderboard
11 Kimi K2 Sep 2025 31.6 Sep 2025 Official leaderboard
12 Gemini 2.5 Flash Sep 2025 31.6 Sep 2025 Official leaderboard
13 GLM-4.5 29.3 Aug 2025 Official leaderboard
14 gpt-oss-120b 28.0 Aug 2025 Official leaderboard
15 Qwen2.5-Coder-32B-Instruct 27.2 Nov 2024 Official leaderboard
16 Qwen3-32B 27.2 Apr 2025 Official leaderboard
17 Qwen3-14B 24.1 Apr 2025 Official leaderboard
18 Qwen2.5-VL-72B 23.7 Jan 2025 Official leaderboard
19 Qwen2.5-Coder-14B-Instruct 22.5 Nov 2024 Official leaderboard
20 Qwen3-8B 20.0 Apr 2025 Official leaderboard
21 Llama 3.1-70B 18.7 Jul 2024 Official leaderboard
22 gpt-oss-20b 15.2 Aug 2025 Official leaderboard
23 Qwen3-4B 14.7 Apr 2025 Official leaderboard
24 Qwen2.5-Coder (7B) 12.4 Sep 2024 Official leaderboard
25 Llama 3.1-8B 11.3 Jul 2024 Official leaderboard
26 Qwen2.5-VL-7B 7.47 Jan 2025 Official leaderboard
27 Qwen3-1.7B 6.53 Apr 2025 Official leaderboard

Rankings by part

We read 4 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as PFT Overall, VQT Action and VQT CLIP. Claude Sonnet 4 leads 2 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

PFT Overall

27 models we track have a score here.

  1. 1 Claude Sonnet 4 41.5%
  2. 2 Claude Opus 4 40.3%
  3. 3 GPT-5 39.5%
  4. 4 o4-mini 37.3%
  5. 5 GPT-4.1 37.1%
  6. 6 o3 34.9%
  7. 7 DeepSeek-R1 May 2025 33.9%
  8. 8 Qwen3-235B-A22B-Instruct Jul 2025 33.3%
  9. 9 Intern-S1 31.9%
  10. 10 DeepSeek-V3 Mar 2025 31.7%

Number one over time

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

Anthropic
Own leaderboard Not tested since Oct 2025

Source: Official InteractScience 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
0.00 10.0 20.0 30.0 40.0 50.0 2025 0.00 10.0 20.0 30.0 40.0 50.0 2025

Source: Official InteractScience 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 Coding

All 523 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Fable 5.2 96.6%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • Codeforces rating In use Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance. DeepSeek V4.1 Flash 3,471

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash. 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 (30 entries).
  • Pass Models scored 27 models we track, plus 3 other entries on its leaderboard.
  • Warning Newest score The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash.
  • 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 github.com: 27 rows for models we track and 3 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