Benchmark Almanac

Mini Amusement Parks

Neuro-SAN-Studio 1861776.0 tops its leaderboard GPT-5 371375.0 leads the models we track
Not in model cards

Agents build rides, hire staff, and manage an amusement park. The simulator scores the park’s final value.

The leaderboard separates park layouts, difficulty levels, and access to instructions. It reports human scores and the cost of agent runs.

Made by
Skyfall AI
Released
November 2025
Kind
Video games
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Actions
Interaction
Agentic
Environment
Mini Amusement Parks
Items from
Generated
Refresh
One fixed set
Run by
Anyone
In views
Games, Agents
Strongest model tested
GPT-5 #71 on the capability index
Last model added
September 2025
Scores refreshed
November 20th, 2026
6
Entries on its own leaderboard, 5 models we track

Neuro-SAN-Studio tops the leaderboard with 1861776.0 GPT-5 leads with 371375.0 among models released by September 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. The best model we track is GPT-5, at 371375.0. 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 Neuro-SAN-Studio the_islands, medium, unlimited ✓ 1861776.0
  2. 2 GPT-5 ribs, easy, few-shot+docs ✓ 371375.0
  3. 3 Grok 4 the_islands, easy, docs ✓ 341999.0
  4. 4 Gemini 2.5 Pro Jun 2025, the_islands, easy, docs ✓ 194298.0
  5. 5 Claude Sonnet 4.5 the_islands, easy, docs ✓ 117568.0
  6. 6 GPT-5 nano ribs, medium, docs ✓ 21806.0

Source: Official Mini Amusement Parks leaderboard

#EntryScoreSettingDateLinkSource
1 Neuro-SAN-Studio ✓ 1861776.0 the_islands, medium, unlimited github.com Official leaderboard
2 GPT-5 ✓ 371375.0 ribs, easy, few-shot+docs Aug 2025 openreview.net Official leaderboard
3 Grok 4 ✓ 341999.0 the_islands, easy, docs Jul 2025 openreview.net Official leaderboard
4 Gemini 2.5 Pro Jun 2025 ✓ 194298.0 the_islands, easy, docs Jun 2025 openreview.net Official leaderboard
5 Claude Sonnet 4.5 ✓ 117568.0 the_islands, easy, docs Sep 2025 openreview.net Official leaderboard
6 GPT-5 nano ✓ 21806.0 ribs, medium, docs Aug 2025 openreview.net Official leaderboard
#ModelScoreSettingDateLinkSource
1 GPT-5 ✓ 371375.0 ribs, easy, few-shot+docs Aug 2025 openreview.net Official leaderboard
2 Grok 4 ✓ 341999.0 the_islands, easy, docs Jul 2025 openreview.net Official leaderboard
3 Gemini 2.5 Pro Jun 2025 ✓ 194298.0 the_islands, easy, docs Jun 2025 openreview.net Official leaderboard
4 Claude Sonnet 4.5 ✓ 117568.0 the_islands, easy, docs Sep 2025 openreview.net Official leaderboard
5 GPT-5 nano ✓ 21806.0 ribs, medium, docs Aug 2025 openreview.net Official leaderboard

The best score, 371375.0, costs $4.15 for the full run

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Google DeepMind OpenAI Cost frontier One model at each effort level
0.00 100000.0 200000.0 300000.0 400000.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-5, ribs, easy, few-shot+docs GPT-5, ribs, easy, docs Gemini 2.5 Pro, the_islands, easy, docs Gemini 2.5 Pro, the_islands, easy, few-shot+docs GPT-5 nano, zig_zag, easy, docs 0.00 100000.0 200000.0 300000.0 400000.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-5, ribs, easy, few-shot+docs GPT-5, ribs, easy, docs

Source: Official Mini Amusement Parks leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-5 ribs, easy, few-shot+docs 371375.0 $4.15
Grok 4 the_islands, easy, docs 341999.0 $7.62
GPT-5 the_islands, medium, few-shot+docs 307490.0 $10
GPT-5 the_islands, easy, few-shot+docs 305408.0 $4.43
GPT-5 ribs, easy, docs 301028.0 $3.67
GPT-5 ribs, medium, docs 260716.0 $8.16
GPT-5 the_islands, easy, docs 224637.0 $4.25
Gemini 2.5 Pro Jun 2025 the_islands, easy, docs 194298.0 $3.61
Claude Sonnet 4.5 the_islands, easy, docs 117568.0 $7.20
GPT-5 zig_zag, easy, docs 113247.0 $3.78
GPT-5 zig_zag, medium, few-shot+docs 62664.0 $9.26
Grok 4 the_islands, easy, few-shot+docs 58569.0 $8.18
Claude Sonnet 4.5 the_islands, medium, few-shot+docs 52685.0 $16
GPT-5 zig_zag, easy, few-shot+docs 51132.0 $4.10
Claude Sonnet 4.5 ribs, medium, docs 48291.0 $13
Grok 4 ribs, medium, few-shot+docs 37848.0 $14
Claude Sonnet 4.5 the_islands, medium, docs 37117.0 $14
Gemini 2.5 Pro Jun 2025 the_islands, easy, few-shot+docs 36703.0 $3.55
GPT-5 zig_zag, medium, docs 32960.0 $8.08
Gemini 2.5 Pro Jun 2025 the_islands, medium, docs 32824.0 $6.51
Grok 4 the_islands, medium, docs 26243.0 $14
Claude Sonnet 4.5 ribs, medium, few-shot+docs 25883.0 $14
Gemini 2.5 Pro Jun 2025 zig_zag, easy, docs 24658.0 $3.30
GPT-5 nano ribs, medium, docs 21806.0 $0.32
Grok 4 the_islands, medium, few-shot+docs 21094.0 $16
Grok 4 ribs, easy, few-shot+docs 20473.0 $7.22
Claude Sonnet 4.5 the_islands, easy, few-shot+docs 19636.0 $8.12
Claude Sonnet 4.5 zig_zag, easy, docs 19302.0 $5.95
Gemini 2.5 Pro Jun 2025 ribs, easy, few-shot+docs 14558.0 $3.10
Grok 4 ribs, medium, docs 13488.0 $13
Grok 4 zig_zag, medium, few-shot+docs 11188.0 $14
Grok 4 zig_zag, easy, few-shot+docs 10864.0 $6.87
Gemini 2.5 Pro Jun 2025 ribs, medium, docs 10670.0 $5.48
Grok 4 zig_zag, easy, docs 8921.0 $6.31
Gemini 2.5 Pro Jun 2025 zig_zag, easy, few-shot+docs 7599.0 $3.23
Claude Sonnet 4.5 zig_zag, easy, few-shot+docs 6743.0 $7.10
Gemini 2.5 Pro Jun 2025 ribs, easy, docs 4867.0 $2.99
GPT-5 nano zig_zag, easy, docs 4182.0 $0.14
GPT-5 nano the_islands, easy, docs 4091.0 $0.16
GPT-5 nano ribs, medium, few-shot+docs 3837.0 $0.32
Grok 4 zig_zag, medium, docs 3023.0 $12
GPT-5 nano ribs, easy, docs 2981.0 $0.13
Claude Sonnet 4.5 ribs, easy, few-shot+docs 2875.0 $6.90
GPT-5 nano ribs, easy, few-shot+docs 2226.0 $0.17
GPT-5 nano zig_zag, medium, few-shot+docs 1740.0 $0.29
GPT-5 nano the_islands, medium, few-shot+docs 1657.0 $0.34

The best score, 371375.0, costs $4.15 for the full run

Runs on the cost frontier score higher than any run that's cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Google DeepMind OpenAI Cost frontier One model at each effort level
0.00 100000.0 200000.0 300000.0 400000.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-5, ribs, easy, few-shot+docs GPT-5, ribs, easy, docs Gemini 2.5 Pro, the_islands, easy, docs Gemini 2.5 Pro, the_islands, easy, few-shot+docs GPT-5 nano, zig_zag, easy, docs 0.00 100000.0 200000.0 300000.0 400000.0 $0.10 $1.00 $10 $100 Cost of the full run, log scale GPT-5, ribs, easy, few-shot+docs GPT-5, ribs, easy, docs

Source: Official Mini Amusement Parks leaderboard

ModelEffortHarnessScoreCost of the full runOutput tokens
GPT-5 ribs, easy, few-shot+docs 371375.0 $4.15
Grok 4 the_islands, easy, docs 341999.0 $7.62
GPT-5 the_islands, medium, few-shot+docs 307490.0 $10
GPT-5 the_islands, easy, few-shot+docs 305408.0 $4.43
GPT-5 ribs, easy, docs 301028.0 $3.67
GPT-5 ribs, medium, docs 260716.0 $8.16
GPT-5 the_islands, easy, docs 224637.0 $4.25
Gemini 2.5 Pro Jun 2025 the_islands, easy, docs 194298.0 $3.61
Claude Sonnet 4.5 the_islands, easy, docs 117568.0 $7.20
GPT-5 zig_zag, easy, docs 113247.0 $3.78
GPT-5 zig_zag, medium, few-shot+docs 62664.0 $9.26
Grok 4 the_islands, easy, few-shot+docs 58569.0 $8.18
Claude Sonnet 4.5 the_islands, medium, few-shot+docs 52685.0 $16
GPT-5 zig_zag, easy, few-shot+docs 51132.0 $4.10
Claude Sonnet 4.5 ribs, medium, docs 48291.0 $13
Grok 4 ribs, medium, few-shot+docs 37848.0 $14
Claude Sonnet 4.5 the_islands, medium, docs 37117.0 $14
Gemini 2.5 Pro Jun 2025 the_islands, easy, few-shot+docs 36703.0 $3.55
GPT-5 zig_zag, medium, docs 32960.0 $8.08
Gemini 2.5 Pro Jun 2025 the_islands, medium, docs 32824.0 $6.51
Grok 4 the_islands, medium, docs 26243.0 $14
Claude Sonnet 4.5 ribs, medium, few-shot+docs 25883.0 $14
Gemini 2.5 Pro Jun 2025 zig_zag, easy, docs 24658.0 $3.30
GPT-5 nano ribs, medium, docs 21806.0 $0.32
Grok 4 the_islands, medium, few-shot+docs 21094.0 $16
Grok 4 ribs, easy, few-shot+docs 20473.0 $7.22
Claude Sonnet 4.5 the_islands, easy, few-shot+docs 19636.0 $8.12
Claude Sonnet 4.5 zig_zag, easy, docs 19302.0 $5.95
Gemini 2.5 Pro Jun 2025 ribs, easy, few-shot+docs 14558.0 $3.10
Grok 4 ribs, medium, docs 13488.0 $13
Grok 4 zig_zag, medium, few-shot+docs 11188.0 $14
Grok 4 zig_zag, easy, few-shot+docs 10864.0 $6.87
Gemini 2.5 Pro Jun 2025 ribs, medium, docs 10670.0 $5.48
Grok 4 zig_zag, easy, docs 8921.0 $6.31
Gemini 2.5 Pro Jun 2025 zig_zag, easy, few-shot+docs 7599.0 $3.23
Claude Sonnet 4.5 zig_zag, easy, few-shot+docs 6743.0 $7.10
Gemini 2.5 Pro Jun 2025 ribs, easy, docs 4867.0 $2.99
GPT-5 nano zig_zag, easy, docs 4182.0 $0.14
GPT-5 nano the_islands, easy, docs 4091.0 $0.16
GPT-5 nano ribs, medium, few-shot+docs 3837.0 $0.32
Grok 4 zig_zag, medium, docs 3023.0 $12
GPT-5 nano ribs, easy, docs 2981.0 $0.13
Claude Sonnet 4.5 ribs, easy, few-shot+docs 2875.0 $6.90
GPT-5 nano ribs, easy, few-shot+docs 2226.0 $0.17
GPT-5 nano zig_zag, medium, few-shot+docs 1740.0 $0.29
GPT-5 nano the_islands, medium, few-shot+docs 1657.0 $0.34

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 Nov 2025
GPT-5 371375.0

Source: Official Mini Amusement Parks 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 100000.0 200000.0 300000.0 400000.0 2025 0.00 100000.0 200000.0 300000.0 400000.0 2025

Source: Official Mini Amusement Parks 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 Games and long tasks

All 261 in Games and long tasks
  • ARC-AGI-3 In use Interactive game environments where agents must explore, infer goals and rules, and act efficiently without instructions. GPT-6.2 Astra 100%
  • Orak Fading LLM agents play 12 popular video games across most genres through an MCP interface. Each game scores in-game progress or outcomes, and the board ranks models by their average rank across the games. Gemini 2.5 Pro 3.90
  • MineDojo Fading Agents complete open-ended tasks in Minecraft, such as mining blocks and fighting mobs, from a large simulation suite. Programmatic checks score many tasks. Seed2.0 Pro 49%
  • TextQuests Only in system cards Agents play 25 classic interactive fiction games. The score is harmful actions per 100 steps, where lower is better. The board also reports game progress. Llama 4 Scout 7.00
  • GameWorld Only in system cards Multimodal game agents play 34 browser games across 170 tasks. Outcomes are checked from the game state. GPT-6.1 Terra 76.8%
  • Swarm Bench Only in system cards Models orchestrate agent swarms on complex tasks that benefit from coordinated decomposition and parallel execution. Kimi K3 76.3%

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 Claude Sonnet 4.5. 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 (6 entries).
  • Pass Models scored 5 models we track, plus 1 other entry on its leaderboard.
  • Warning Newest score The newest came 14 months ago, in September 2025, for Claude Sonnet 4.5.
  • 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 maps.skyfall.ai: 46 rows for models we track and 1 other entry 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