Gert Labs Composite Game Benchmark (Gert Labs)
We show this table for reference; we do not rank on it.
A game-environment benchmark that evaluates AI models in novel games covering strategic planning, resource management, spatial reasoning, cooperation, and theory of mind.
Benchmark score on Gert Labs — November 13, 2026
We compile the Gert Labs rows from benchmark-owner or independent runs. Claude Opus 4.8 leads the table at 72.97%, followed by GPT-5.5 (72.93%) and Claude Opus 4.7 (65.59%). We do not use these results to rank models overall.
Claude Opus 4.8
Anthropic
Run by Gert LabsGPT-5.5
OpenAI
Run by Gert LabsClaude Opus 4.7
Anthropic
Run by Gert Labs55 modelsAgenticCurrentDisplay onlyUpdated November 13, 2026
Benchmark score table (55 models)
| Rank | Model / configuration | Score | Who ran it | Parameters (B) | Open / closed |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.8AnthropicRun by Gert Labs | 72.97% | Run by Gert Labs | Not reported | Closed |
| 2 | GPT-5.5OpenAIRun by Gert Labs | 72.93% | Run by Gert Labs | Not reported | Closed |
| 3 | Claude Opus 4.7AnthropicRun by Gert Labs | 65.59% | Run by Gert Labs | Not reported | Closed |
| 4 | GPT-5.4OpenAIRun by Gert Labs | 64.89% | Run by Gert Labs | Not reported | Closed |
| 5 | Qwen3.7 MaxAlibabaRun by Gert Labs | 64.27% | Run by Gert Labs | Not reported | Closed |
| 6 | Claude Opus 4.5AnthropicRun by Gert Labs | 64.23% | Run by Gert Labs | Not reported | Closed |
| 7 | Gemini 3 ProGoogleRun by Gert Labs | 63.23% | Run by Gert Labs | Not reported | Closed |
| 8 | Claude Sonnet 4.6AnthropicRun by Gert Labs | 62.92% | Run by Gert Labs | Not reported | Closed |
| 9 | MiMo-V2.5-ProXiaomiRun by Gert Labs | 62.70% | Run by Gert Labs | Not reported | Closed |
| 10 | Gemini 3.5 FlashGoogleRun by Gert Labs | 61.85% | Run by Gert Labs | Not reported | Closed |
| 11 | Claude Opus 4.6AnthropicRun by Gert Labs | 61.85% | Run by Gert Labs | Not reported | Closed |
| 12 | GLM-5.1Z.AIRun by Gert Labs | 60.11% | Run by Gert Labs | Not reported | Open |
| 13 | GPT-5.3 CodexOpenAIRun by Gert Labs | 57.47% | Run by Gert Labs | Not reported | Closed |
| 14 | Gemini 3.1 ProGoogleRun by Gert Labs | 56.87% | Run by Gert Labs | Not reported | Closed |
| 15 | Kimi K2.6Moonshot AIRun by Gert Labs | 56.82% | Run by Gert Labs | Not reported | Open |
| 16 | Gemini 3 FlashGoogleRun by Gert Labs | 56.63% | Run by Gert Labs | Not reported | Closed |
| 17 | Qwen3.6-27BAlibabaRun by Gert Labs | 54.84% | Run by Gert Labs | Not reported | Open |
| 18 | DeepSeek V4 FlashDeepSeekRun by Gert Labs | 54.35% | Run by Gert Labs | Not reported | Open |
| 19 | GPT-5.2-CodexOpenAIRun by Gert Labs | 51.79% | Run by Gert Labs | Not reported | Closed |
| 20 | Step 3.7 FlashStepFunRun by Gert Labs | 51.57% | Run by Gert Labs | Not reported | Open |
| 21 | GLM-5Z.AIRun by Gert Labs | 50.99% | Run by Gert Labs | Not reported | Open |
| 22 | Qwen3.6 PlusAlibabaRun by Gert Labs | 50.60% | Run by Gert Labs | Not reported | Closed |
| 23 | GPT-5.1-CodexOpenAIRun by Gert Labs | 49.68% | Run by Gert Labs | Not reported | Closed |
| 24 | Grok Build 0.1xAIRun by Gert Labs | 49.15% | Run by Gert Labs | Not reported | Closed |
| 25 | Claude Sonnet 4.5AnthropicRun by Gert Labs | 48.51% | Run by Gert Labs | Not reported | Closed |
| 26 | Grok 4.1 FastxAIRun by Gert Labs | 47.32% | Run by Gert Labs | Not reported | Closed |
| 27 | MiMo-V2.5XiaomiRun by Gert Labs | 46.89% | Run by Gert Labs | Not reported | Closed |
| 28 | Qwen3.5 397BAlibabaRun by Gert Labs | 46.76% | Run by Gert Labs | Not reported | Open |
| 29 | GPT-5.2OpenAIRun by Gert Labs | 46.54% | Run by Gert Labs | Not reported | Closed |
| 30 | Kimi K2.5Moonshot AIRun by Gert Labs | 45.88% | Run by Gert Labs | Not reported | Open |
| 31 | Grok 4.3xAIRun by Gert Labs | 43.86% | Run by Gert Labs | Not reported | Closed |
| 32 | Qwen3 MaxAlibabaRun by Gert Labs | 43.74% | Run by Gert Labs | Not reported | Closed |
| 33 | Qwen3.6-35B-A3BAlibabaRun by Gert Labs | 42.65% | Run by Gert Labs | Not reported | Open |
| 34 | Grok 4xAIRun by Gert Labs | 42.34% | Run by Gert Labs | Not reported | Closed |
| 35 | Gemini 2.5 ProGoogleRun by Gert Labs | 42.01% | Run by Gert Labs | Not reported | Closed |
| 36 | GPT-5.1OpenAIRun by Gert Labs | 41.24% | Run by Gert Labs | Not reported | Closed |
| 37 | MiniMax M2.7MiniMaxRun by Gert Labs | 40.40% | Run by Gert Labs | Not reported | Open |
| 38 | GLM-4.7Z.AIRun by Gert Labs | 39.95% | Run by Gert Labs | Not reported | Open |
| 39 | Claude 4 SonnetAnthropicRun by Gert Labs | 39.66% | Run by Gert Labs | Not reported | Closed |
| 40 | Qwen3.5-27BAlibabaRun by Gert Labs | 39.41% | Run by Gert Labs | Not reported | Open |
| 41 | MiniMax M2.5MiniMaxRun by Gert Labs | 39.11% | Run by Gert Labs | Not reported | Closed |
| 42 | Mistral Medium 3.5 128BMistralRun by Gert Labs | 39.10% | Run by Gert Labs | Not reported | Open |
| 43 | Gemini 3.1 Flash-LiteGoogleRun by Gert Labs | 38.46% | Run by Gert Labs | Not reported | Closed |
| 44 | Grok 4.20xAIRun by Gert Labs | 38.36% | Run by Gert Labs | Not reported | Closed |
| 45 | Hy3 PreviewTencentRun by Gert Labs | 36.91% | Run by Gert Labs | Not reported | Open |
| 46 | MiMo-V2-ProXiaomiRun by Gert Labs | 36.68% | Run by Gert Labs | Not reported | Closed |
| 47 | Gemma 4 31BGoogleRun by Gert Labs | 35.26% | Run by Gert Labs | Not reported | Open |
| 48 | Kimi K2.5 (Reasoning)Moonshot AIRun by Gert Labs | 32.58% | Run by Gert Labs | Not reported | Open |
| 49 | Trinity-Large-ThinkingArcee AIRun by Gert Labs | 32.55% | Run by Gert Labs | Not reported | Open |
| 50 | GLM-5V-TurboZ.AIRun by Gert Labs | 30.76% | Run by Gert Labs | Not reported | Closed |
| 51 | GPT-OSS 120BOpenAIRun by Gert Labs | 29.61% | Run by Gert Labs | Not reported | Open |
| 52 | DeepSeek V3.2DeepSeekRun by Gert Labs | 29.57% | Run by Gert Labs | Not reported | Open |
| 53 | Qwen3.5-35B-A3BAlibabaRun by Gert Labs | 28.96% | Run by Gert Labs | Not reported | Open |
| 54 | GPT-4.1OpenAIRun by Gert Labs | 25.65% | Run by Gert Labs | Not reported | Closed |
| 55 | Nemotron 3 Super 120B A12BNVIDIARun by Gert Labs | 25.34% | Run by Gert Labs | Not reported | Open |
- Lab-reported
- Benchmark-run
- Independent
Among the reported Gert Labs rows, Claude Opus 4.8 is first at 72.97%. The third row is 7.38 points behind. The broader top-10 range is 11.12 points, so the table still separates the published systems.
55 models have been evaluated on Gert Labs. The benchmark falls in the Agentic category. Gert Labs is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Gert Labs
Year
2026
Tasks
Novel game environments
Format
Composite game leaderboard
Difficulty
Agentic coding and decision-making
The public Gert Labs leaderboard reports a composite 0-100 metric derived from average and median percentile across games, success rate, and response-time penalty. The combined leaderboard blends agentic coding, one-shot coding, and social decision-making modes.
Freshness and provenance
Version
Gert Labs 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Gert Labs measure?
A game-environment benchmark that evaluates AI models in novel games covering strategic planning, resource management, spatial reasoning, cooperation, and theory of mind.
Which model scores highest on Gert Labs?
Claude Opus 4.8 by Anthropic currently leads with a score of 72.97% on Gert Labs.
How many models are evaluated on Gert Labs?
55 AI models have published results on Gert Labs in the BenchLM catalog.
Compare top models on Gert Labs
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.