Local LLM Arkanoid Test
Agents build a mouse-controlled brick breaker in one HTML file. They can revise it before a separate Claude agent checks 11 requirements in a browser.
The maker publishes 13 scored runs covering 11 distinct models. The board repeats one model through a second provider and another at a second quantization.
- Made by
- AOX.cz
- Released
- May 2026
- Kind
- Web and app front ends
- Level
- University
- Scored
- Graded by a model
- Answers
- Code
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Multi-turn
- Items from
- Written by experts
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- GPT-5.5 #21 on the capability index
- Last model added
- April 2026
- Scores refreshed
- November 20th, 2026
9 models share the top score of 11.0 among models released by April 2026 9 models share the top score of 11.0 among models released by April 2026
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. 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 April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 openrouter/google/gemini-3-flash-preview 11.0
- 2 gpt-oss-20b mxfp4 11.0
- 3
Gemini 2.5 Flash Sep 2025 11.0
- 4
Gemini 3 Flash 11.0
- 5
Gemini 3.1 Pro 11.0
- 6
Qwen3.5-9B q4_k_m 11.0
- 7
Gemini 3.1 Flash-Lite 11.0
- 8
Gemma 4 E2B q8_0 11.0
- 9
Gemma 4 E4B q8_0 11.0
- 10 GPT-5.5 11.0
- 11 openrouter/google/gemini-3.5-flash 10.0
- 12
Nemotron 3 Nano 4B q8_0 10.0
Source: Official Local LLM Arkanoid Test leaderboard
- 1 gpt-oss-20b mxfp4 11.0
- 2
Gemini 2.5 Flash Sep 2025 11.0
- 3
Gemini 3 Flash 11.0
- 4
Gemini 3.1 Pro 11.0
- 5
Qwen3.5-9B q4_k_m 11.0
- 6
Gemini 3.1 Flash-Lite 11.0
- 7
Gemma 4 E2B q8_0 11.0
- 8
Gemma 4 E4B q8_0 11.0
- 9 GPT-5.5 11.0
- 10
Nemotron 3 Nano 4B q8_0 10.0
Source: Official Local LLM Arkanoid Test leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | openrouter/google/gemini-3-flash-preview | 11.0 | aox.cz | Official leaderboard | ||
| 2 | gpt-oss-20b | 11.0 | mxfp4 | Aug 2025 | aox.cz | Official leaderboard |
| 3 | | 11.0 | Sep 2025 | aox.cz | Official leaderboard | |
| 4 | | 11.0 | Dec 2025 | aox.cz | Official leaderboard | |
| 5 | | 11.0 | Feb 2026 | aox.cz | Official leaderboard | |
| 6 | | 11.0 | q4_k_m | Feb 2026 | aox.cz | Official leaderboard |
| 7 | | 11.0 | Mar 2026 | aox.cz | Official leaderboard | |
| 8 | | 11.0 | q8_0 | Mar 2026 | aox.cz | Official leaderboard |
| 9 | | 11.0 | q8_0 | Mar 2026 | aox.cz | Official leaderboard |
| 10 | GPT-5.5 | 11.0 | Apr 2026 | aox.cz | Official leaderboard | |
| 11 | openrouter/google/gemini-3.5-flash | 10.0 | aox.cz | Official leaderboard | ||
| 12 | | 10.0 | q8_0 | Mar 2026 | aox.cz | Official leaderboard |
| # | Model | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | gpt-oss-20b | 11.0 | mxfp4 | Aug 2025 | aox.cz | Official leaderboard |
| 2 | | 11.0 | Sep 2025 | aox.cz | Official leaderboard | |
| 3 | | 11.0 | Dec 2025 | aox.cz | Official leaderboard | |
| 4 | | 11.0 | Feb 2026 | aox.cz | Official leaderboard | |
| 5 | | 11.0 | q4_k_m | Feb 2026 | aox.cz | Official leaderboard |
| 6 | | 11.0 | Mar 2026 | aox.cz | Official leaderboard | |
| 7 | | 11.0 | q8_0 | Mar 2026 | aox.cz | Official leaderboard |
| 8 | | 11.0 | q8_0 | Mar 2026 | aox.cz | Official leaderboard |
| 9 | GPT-5.5 | 11.0 | Apr 2026 | aox.cz | Official leaderboard | |
| 10 | | 10.0 | q8_0 | Mar 2026 | aox.cz | Official leaderboard |
Rankings by part
We read 11 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Loads cleanly, Mouse paddle control and Ball launches. gpt-oss-20b leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
Loads cleanly
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Mouse paddle control
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Ball launches
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Wall bounce
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Paddle bounce
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Brick field
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Destroy brick + score
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Paddle deflection angle
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Life loss
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Game over + restart
10 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Nemotron 3 Nano 4B ↑3 1.00
- 8
Gemma 4 E2B ↓1 1.00
- 9
Gemma 4 E4B ↓1 1.00
- 10 GPT-5.5 ↓1 1.00
Level clear / win
9 models we track have a score here.
- 1 gpt-oss-20b 1.00
- 2
Gemini 2.5 Flash Sep 2025 1.00
- 3
Gemini 3 Flash 1.00
- 4
Gemini 3.1 Pro 1.00
- 5
Qwen3.5-9B 1.00
- 6
Gemini 3.1 Flash-Lite 1.00
- 7
Gemma 4 E2B 1.00
- 8
Gemma 4 E4B 1.00
- 9 GPT-5.5 1.00
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Local LLM Arkanoid Test leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Local LLM Arkanoid Test 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
- 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 7 months ago, in April 2026, for GPT-5.5. 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 (12 entries).
- Pass Models scored 10 models we track, plus 2 other entries on its leaderboard.
- Warning Newest score The newest came 7 months ago, in April 2026, for GPT-5.5.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-5.5 (#21).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on aox.cz: 11 rows for models we track and 2 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.