Skip to content

Benchmark

ExplorationBench · AlienLogic · M4 · fixed-probe control

Scores are percentages. Higher is better.

Models and systems measured
10
Scores recorded
10
Sources
1
IQ weight
No weight published

Score distribution

51.9%
Mean
54.3%
Median
5.8%
Std dev
43.8%
10th percentile
58.6%
90th percentile
41%
Min
41%59%
The band covers the middle 80% of scores. The line marks the median.

Results

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Claude Opus 5official 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · max reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs48.6%
DeepSeek V4 Pro (version unspecified)official 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · max reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs43.8%
DeepSeek V4.1 Flashofficial 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · max reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs41%
Gemini 3.8 Flashofficial 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · high reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs57.6%
GPT-5.6 Solofficial 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · max reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs58.6%
Grok 4.6official 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · xhigh reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs54.3%
Hy4 previewofficial 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · high reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.4%
Kimi K3official 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · max reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs54.3%
Qwen3.8 Max 0902official 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · max reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs59%
Seed2.1 Pro 0915official 2026-10-13 leaderboard snapshot · 70 held-out tasks in this world · 4 exploration rounds · 3 answers/task · high reasoning · one matched control trajectory· Reported by Fudan University / Tencent Hy / Tsinghua University· via BenchBoard docs50.5%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.