Benchmark Almanac

Crest Arcade Wildhaven

Gemma 3 4B 100% leads since March 2025
Not in model cards

Models select a team of animals and choose where each lands in a village. The score measures the share of village strength they destroy.

Glimmer Index publishes scores for 15 distinct models. Of those, 9 completed the required set.

Made by
Glimmer Index
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
Wildhaven
Items from
Generated
Refresh
One fixed set
Run by
Anyone
In views
Games, Agents
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
15
Entries on its own leaderboard, 14 models we track

Gemma 3 4B leads with 100% Gemma 3 4B leads with 100%

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. This view shows only models we track, each at its best.

  1. 1 Gemma 3 4B none · Provisional 100%
  2. 2 Gemini 4 Neon medium · Provisional 91%
  3. 3 GPT-6.2 Astra medium 85.6%
  4. 4 GPT-6 Astra medium 84.4%
  5. 5 GPT-6.1 Sol medium 83.6%
  6. 6 Gemini 4 Hydrogen medium 82.6%
  7. 7 GPT-6 Sol medium 80.4%
  8. 8 GPT-6.1 Terra medium 78.1%
  9. 9 Claude Sonnet 5.5 medium 71.9%
  10. 10 Qwen3.8-Flash-Next low · Provisional 71.4%
  11. 11 Qwen 3.8 27B none · Provisional 66.7%
  12. 12 Claude Haiku 5.5 medium 66%
  13. 13 GPT-6 Luna medium 62.3%
  14. 14 Qwen3.5-9B none · Provisional 60.7%
  15. 15 Ministral 3 8B none · Provisional 57.7%

Source: Official Crest Arcade Wildhaven leaderboard

#EntryScore ±SettingReleasedRun onSource
1 Gemma 3 4B 100% 2 to 100 none · Provisional Mar 2025 Nov 2026 Official leaderboard
2 Gemini 4 Neon 91% 42 to 100 medium · Provisional Nov 2026 Nov 2026 Official leaderboard
3 GPT-6.2 Astra 85.6% ±5.3 medium Nov 2026 Nov 2026 Official leaderboard
4 GPT-6 Astra 84.4% ±5.2 medium Sep 2026 Oct 2026 Official leaderboard
5 GPT-6.1 Sol 83.6% ±5.2 medium Sep 2026 Oct 2026 Official leaderboard
6 Gemini 4 Hydrogen 82.6% ±5.6 medium Nov 2026 Nov 2026 Official leaderboard
7 GPT-6 Sol 80.4% ±5.7 medium Sep 2026 Oct 2026 Official leaderboard
8 GPT-6.1 Terra 78.1% ±6 medium Oct 2026 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 71.9% ±6.3 medium Sep 2026 Nov 2026 Official leaderboard
10 Qwen3.8-Flash-Next 71.4% ±23.3 low · Provisional Aug 2026 Nov 2026 Official leaderboard
11 Qwen 3.8 27B 66.7% 10.1 to 100 none · Provisional Aug 2026 Nov 2026 Official leaderboard
12 Claude Haiku 5.5 66% ±6.4 medium Oct 2026 Nov 2026 Official leaderboard
13 GPT-6 Luna 62.3% ±6.2 medium Sep 2026 Oct 2026 Official leaderboard
14 Qwen3.5-9B 60.7% 4.1 to 100 none · Provisional Feb 2026 Nov 2026 Official leaderboard
15 Ministral 3 8B open weights 57.7% 1.1 to 100 none · Provisional Nov 2026 Official leaderboard
#ModelScore ±SettingReleasedRun onSource
1 Gemma 3 4B 100% 2 to 100 none · Provisional Mar 2025 Nov 2026 Official leaderboard
2 Gemini 4 Neon 91% 42 to 100 medium · Provisional Nov 2026 Nov 2026 Official leaderboard
3 GPT-6.2 Astra 85.6% ±5.3 medium Nov 2026 Nov 2026 Official leaderboard
4 GPT-6 Astra 84.4% ±5.2 medium Sep 2026 Oct 2026 Official leaderboard
5 GPT-6.1 Sol 83.6% ±5.2 medium Sep 2026 Oct 2026 Official leaderboard
6 Gemini 4 Hydrogen 82.6% ±5.6 medium Nov 2026 Nov 2026 Official leaderboard
7 GPT-6 Sol 80.4% ±5.7 medium Sep 2026 Oct 2026 Official leaderboard
8 GPT-6.1 Terra 78.1% ±6 medium Oct 2026 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 71.9% ±6.3 medium Sep 2026 Nov 2026 Official leaderboard
10 Qwen3.8-Flash-Next 71.4% ±23.3 low · Provisional Aug 2026 Nov 2026 Official leaderboard
11 Qwen 3.8 27B 66.7% 10.1 to 100 none · Provisional Aug 2026 Nov 2026 Official leaderboard
12 Claude Haiku 5.5 66% ±6.4 medium Oct 2026 Nov 2026 Official leaderboard
13 GPT-6 Luna 62.3% ±6.2 medium Sep 2026 Oct 2026 Official leaderboard
14 Qwen3.5-9B 60.7% 4.1 to 100 none · Provisional Feb 2026 Nov 2026 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s own leaderboard: Hunt, Reckoning and Outmatched. Each ranks models on one part of the test. Gemini 4 Neon leads 2 of them, and Gemma 3 4B leads the other one. Arrows show how far a model moved from its overall rank among the models we track.

Hunt

14 models we track have a score here.

  1. 1 Gemma 3 4B 100%
  2. 2 Qwen3.5-9B ↑12 100%
  3. 3 Qwen 3.8 27B ↑8 100%
  4. 4 Gemini 4 Neon ↓2 100%
  5. 5 GPT-6 Astra ↓1 98.1%
  6. 6 GPT-6.1 Sol ↓1 97.3%
  7. 7 GPT-6.2 Astra ↓4 96.8%
  8. 8 Gemini 4 Hydrogen ↓2 96.1%
  9. 9 GPT-6.1 Terra ↓1 94.9%
  10. 10 Qwen3.8-Flash-Next 93.3%

Of runs with a cost, the best scores 91% for $0.41 a task

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.

OpenAI Anthropic Google DeepMind Cost frontier One model at each effort level
0% 20% 40% 60% 80% 100% $0.001 $0.01 $0.10 $1.00 Cost per task, log scale Gemini 4 Neon, medium · Provisional GPT-6.2 Astra, medium GPT-6.1 Sol, medium GPT-6.1 Terra, medium Claude Haiku 5.5, medium 0% 20% 40% 60% 80% 100% $0.001 $0.01 $0.10 $1.00 Cost per task, log scale Gemini 4 Neon, medium · Provisional GPT-6.2 Astra, medium

Source: Official Crest Arcade Wildhaven leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemma 3 4B none · Provisional 100%
Gemini 4 Neon medium · Provisional 91% $0.41
GPT-6.2 Astra medium 85.6% $0.10
GPT-6 Astra medium 84.4% $0.16
GPT-6.1 Sol medium 83.6% $0.04
Gemini 4 Hydrogen medium 82.6% $0.11
GPT-6 Sol medium 80.4% $0.04
GPT-6.1 Terra medium 78.1% $0.01
Claude Sonnet 5.5 medium 71.9% $0.04
Qwen3.8-Flash-Next low · Provisional 71.4%
Qwen 3.8 27B none · Provisional 66.7%
Claude Haiku 5.5 medium 66% $0.004
GPT-6 Luna medium 62.3% $0.002
Qwen3.5-9B none · Provisional 60.7%

Of runs with a cost, the best scores 91% for $0.41 a task

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.

OpenAI Anthropic Google DeepMind Cost frontier One model at each effort level
0% 20% 40% 60% 80% 100% $0.001 $0.01 $0.10 $1.00 Cost per task, log scale Gemini 4 Neon, medium · Provisional GPT-6.2 Astra, medium GPT-6.1 Sol, medium GPT-6.1 Terra, medium Claude Haiku 5.5, medium 0% 20% 40% 60% 80% 100% $0.001 $0.01 $0.10 $1.00 Cost per task, log scale Gemini 4 Neon, medium · Provisional GPT-6.2 Astra, medium

Source: Official Crest Arcade Wildhaven leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemma 3 4B none · Provisional 100%
Gemini 4 Neon medium · Provisional 91% $0.41
GPT-6.2 Astra medium 85.6% $0.10
GPT-6 Astra medium 84.4% $0.16
GPT-6.1 Sol medium 83.6% $0.04
Gemini 4 Hydrogen medium 82.6% $0.11
GPT-6 Sol medium 80.4% $0.04
GPT-6.1 Terra medium 78.1% $0.01
Claude Sonnet 5.5 medium 71.9% $0.04
Qwen3.8-Flash-Next low · Provisional 71.4%
Qwen 3.8 27B none · Provisional 66.7%
Claude Haiku 5.5 medium 66% $0.004
GPT-6 Luna medium 62.3% $0.002
Qwen3.5-9B none · Provisional 60.7%

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 Passed 85% in Mar 2025, by Gemma 3 4B
Gemma 3 4B 100%

Source: Official Crest Arcade Wildhaven leaderboard

Gemma 3 4B on the official leaderboard passed 85% in March 2025

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% 20% 40% 60% 80% 100% 2025 2026 2027 Near the ceiling (85%) Gemma 3 4B 0% 20% 40% 60% 80% 100% 2025 2026 2027 Near the ceiling (85%) Gemma 3 4B

Source: Official Crest Arcade Wildhaven 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

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (15 entries).
  • Pass Models scored 14 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Gemini 4 Neon.
  • Pass Frontier models 5 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Note Headroom Saturated. The best is 100%, past 85% since March 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on glimmerindex.com: 14 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