# Gert Labs Composite Game Benchmark (Gert Labs) > A game-environment benchmark that evaluates AI models in novel games covering strategic planning, resource management, spatial reasoning, cooperation, and theory of mind. Canonical page: https://benchlm.ai/benchmarks/gertlabs - Category: [Agentic](/agentic) - Last updated: September 27, 2026 ## About Gert Labs - Year: 2026 - Tasks: Novel game environments - Format: Composite game leaderboard - Difficulty: Agentic coding and decision-making - Paper: [Gert Labs rankings](https://gertlabs.com/rankings) 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. Gert Labs is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula. ## Leaderboard (52 models) | Rank | Model | Creator | Score | |------|-------|---------|-------| | 1 | [Claude Opus 4.8](/models/claude-opus-4-8) | Anthropic | 72.97% | | 2 | [GPT-5.5](/models/gpt-5-5) | OpenAI | 72.93% | | 3 | [Claude Opus 4.7](/models/claude-opus-4-7) | Anthropic | 65.59% | | 4 | [GPT-5.4](/models/gpt-5-4) | OpenAI | 64.89% | | 5 | [Qwen3.7 Max](/models/qwen3-7-max) | Alibaba | 64.27% | | 6 | [Claude Opus 4.5](/models/claude-opus-4-5) | Anthropic | 64.23% | | 7 | [Gemini 3 Pro](/models/gemini-3-pro) | Google | 63.23% | | 8 | [Claude Sonnet 4.6](/models/claude-sonnet-4-6) | Anthropic | 62.92% | | 9 | [MiMo-V2.5-Pro](/models/mimo-v2-5-pro) | Xiaomi | 62.70% | | 10 | [Claude Opus 4.6](/models/claude-opus-4-6) | Anthropic | 61.85% | | 11 | [Gemini 3.5 Flash](/models/gemini-3-5-flash) | Google | 61.85% | | 12 | [GLM-5.1](/models/glm-5-1) | Z.AI | 60.11% | | 13 | [GPT-5.3 Codex](/models/gpt-5-3-codex) | OpenAI | 57.47% | | 14 | [Gemini 3.1 Pro](/models/gemini-3-1-pro) | Google | 56.87% | | 15 | [Kimi K2.6](/models/kimi-2-6) | Moonshot AI | 56.82% | | 16 | [Gemini 3 Flash](/models/gemini-3-flash) | Google | 56.63% | | 17 | [Qwen3.6-27B](/models/qwen3-6-27b) | Alibaba | 54.84% | | 18 | [GPT-5.2-Codex](/models/gpt-5-2-codex) | OpenAI | 51.79% | | 19 | [Step 3.7 Flash](/models/step-3-7-flash) | StepFun | 51.57% | | 20 | [GLM-5](/models/glm-5) | Z.AI | 50.99% | | 21 | [Qwen3.6 Plus](/models/qwen3-6-plus) | Alibaba | 50.60% | | 22 | [GPT-5.1-Codex](/models/gpt-5-1-codex) | OpenAI | 49.68% | | 23 | [Grok Build 0.1](/models/grok-build-0-1) | xAI | 49.15% | | 24 | [Claude Sonnet 4.5](/models/claude-sonnet-4-5) | Anthropic | 48.51% | | 25 | [Grok 4.1 Fast](/models/grok-4-1-fast) | xAI | 47.32% | | 26 | [MiMo-V2.5](/models/mimo-v2-5) | Xiaomi | 46.89% | | 27 | [Qwen3.5 397B](/models/qwen3-5-397b) | Alibaba | 46.76% | | 28 | [GPT-5.2](/models/gpt-5-2) | OpenAI | 46.54% | | 29 | [Kimi K2.5](/models/kimi-k2-5) | Moonshot AI | 45.88% | | 30 | [Grok 4.3](/models/grok-4-3) | xAI | 43.86% | | 31 | [Qwen3 Max](/models/qwen3-max) | Alibaba | 43.74% | | 32 | [Qwen3.6-35B-A3B](/models/qwen3-6-35b-a3b) | Alibaba | 42.65% | | 33 | [Grok 4](/models/grok-4) | xAI | 42.34% | | 34 | [Gemini 2.5 Pro](/models/gemini-2-5-pro) | Google | 42.01% | | 35 | [GPT-5.1](/models/gpt-5-1) | OpenAI | 41.24% | | 36 | [MiniMax M2.7](/models/minimax-m2-7) | MiniMax | 40.40% | | 37 | [GLM-4.7](/models/glm-4-7) | Z.AI | 39.95% | | 38 | [Claude 4 Sonnet](/models/claude-4-sonnet) | Anthropic | 39.66% | | 39 | [Qwen3.5-27B](/models/qwen3-5-27b) | Alibaba | 39.41% | | 40 | [Mistral Medium 3.5 128B](/models/mistral-medium-3-5-128b) | Mistral | 39.10% | | 41 | [Gemini 3.1 Flash-Lite](/models/gemini-3-1-flash-lite) | Google | 38.46% | | 42 | [Grok 4.20](/models/grok-4-20-beta) | xAI | 38.36% | | 43 | [Hy3 Preview](/models/hy3-preview) | Tencent | 36.91% | | 44 | [MiMo-V2-Pro](/models/mimo-v2-pro) | Xiaomi | 36.68% | | 45 | [Gemma 4 31B](/models/gemma-4-31b) | Google | 35.26% | | 46 | [Kimi K2.5 (Reasoning)](/models/kimi-k2-5-reasoning) | Moonshot AI | 32.58% | | 47 | [Trinity-Large-Thinking](/models/trinity-large-thinking) | Arcee AI | 32.55% | | 48 | [GLM-5V-Turbo](/models/glm-5v-turbo) | Z.AI | 30.76% | | 49 | [GPT-OSS 120B](/models/gpt-oss-120b) | OpenAI | 29.61% | | 50 | [DeepSeek V3.2](/models/deepseek-v3-2) | DeepSeek | 29.57% | | 51 | [Qwen3.5-35B-A3B](/models/qwen3-5-35b-a3b) | Alibaba | 28.96% | | 52 | [GPT-4.1](/models/gpt-4-1) | OpenAI | 25.65% | ## FAQ ### 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? 52 AI models have been evaluated on Gert Labs on BenchLM. ### Does Gert Labs affect BenchLM's overall score? Not directly. Gert Labs is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula. ## Compare Top Models on Gert Labs - [Claude Opus 4.8 vs GPT-5.5](/compare/claude-opus-4-8-vs-gpt-5-5) - [GPT-5.5 vs Claude Opus 4.7](/compare/claude-opus-4-7-vs-gpt-5-5) - [Claude Opus 4.7 vs GPT-5.4](/compare/claude-opus-4-7-vs-gpt-5-4) - [GPT-5.4 vs Qwen3.7 Max](/compare/gpt-5-4-vs-qwen3-7-max)