{
  "schemaVersion": 1,
  "benchmark": "gta",
  "name": {
    "sk": "GTA benchmark LLM modelov",
    "en": "GTA LLM benchmark"
  },
  "url": {
    "sk": "https://dubovskylabs.com/gta-benchmark/",
    "en": "https://dubovskylabs.com/en/gta-benchmark/"
  },
  "dateModified": "2026-08-08",
  "creator": {
    "name": "Marek Dubovsky",
    "url": "https://dubovskylabs.com/"
  },
  "methodology": {
    "sk": "Rovnaký prompt; reálne spustenie výstupu a manuálne praktické hodnotenie hrateľnosti a chýb.",
    "en": "The same prompt; each output was launched and manually evaluated for practical playability and bugs."
  },
  "prompt": {
    "originalLanguage": "en",
    "original": "Create a 3D open-world action game inspired by classic GTA-style gameplay. The game should have a stylized visual style close to GTA: San Andreas, but with improved graphics, better lighting, more detailed environments, and smoother animations. Do not use copyrighted GTA assets, names, characters, maps, or exact copies; create an original game with a similar open-world crime-action feeling.\n\nThe game should use a third-person camera. The player must be able to walk, run, sprint, fight in melee combat, shoot different types of weapons, and interact with vehicles. Weapons should be placed in various harder-to-reach locations across the map, or the player should be able to obtain them from armed gangsters or police officers after defeating them.\n\nThe player must be able to enter cars, steal cars, pull drivers out of vehicles, and drive around the city. The city should be inspired by Miami, with a tropical coastal atmosphere, boats, the sea, canals, bridges, beaches, hotels, highways, and dense urban districts.\n\nThe map should be fairly large and should not be a simple square grid. The road layout should include streets of different shapes, curves, alleys, intersections, multi-level intersections, bridges, tunnels, and a highway network with exits leading into city streets. The city should contain different districts with different architecture, atmosphere, culture, population density, and types of people, similar to real American cities.\n\nInclude districts such as:\n\n* a Chinatown-style district,\n* a Hispanic district,\n* a district with a larger African-American community,\n* a wealthy residential district,\n* a downtown district with high-rise buildings,\n* office districts with business buildings,\n* coastal hotel areas with balconies, pools, bars, tourists, and people relaxing,\n* dangerous gang-controlled districts,\n* and one district occupied by homeless people living in tents.\n\nThe beach areas should be full of life. There should be people swimming, sunbathing, playing beach volleyball, walking, talking, skating, rollerblading, cycling, and relaxing near the sea. Streets in different districts should have different population densities, and pedestrians should visually and behaviorally match the district they are in.\n\nSome dangerous districts should be controlled by gangs. Gangs should be visually recognizable by their clothing, colors, behavior, and preferred weapons. Gang members should patrol their areas, react aggressively to threats, and sometimes attack the player if provoked.\n\nCreate a custom map-generation format and tool for building the city. Think of it like a simplified SimCity-style internal format that describes roads, buildings, parks, parking lots, beaches, canals, bridges, highways, districts, spawn zones, and other city components. Use this format to generate repeated but varied patterns of buildings, roads, parks, parking lots, sidewalks, props, and environmental details. The city should feel handcrafted but should be generated from reusable rules and templates.\n\nOn the edge of the city, include an airport where planes can land and take off. The airport should have runways, hangars, parked vehicles, security areas, and moving aircraft.\n\nThe city should include civilians, gang members, police officers, police cars, police motorcycles, and two police helicopters. Police officers should appear as pedestrians, motorcycle officers, and drivers of police vehicles. The police should only start chasing the player when they witness the player committing a crime, such as shooting in the street, attacking civilians, stealing cars, or causing serious destruction. Police should not be everywhere; they should be distributed naturally so that the player is not immediately chased after the first small incident unless police are nearby and actually notice it.\n\nCivilians should have different personalities and reactions. Most people should panic and run away when the player starts shooting, attacks someone, or tries to steal a car. Some drivers should panic and quickly drive away if the player tries to steal their vehicle but is too slow. A smaller number of civilians should be brave or aggressive and may fight back if the player attacks them, damages their car, or tries to steal their vehicle. However, if the player starts shooting, even these braver civilians should try to escape.\n\nIn dangerous districts, include dealers offering strange substances that can temporarily cause altered visual effects, distorted controls, hallucination-like effects, or other weird states for the player. These districts may also include prostitutes, addicts, and other street characters to make the atmosphere feel darker and more dangerous. Also include bars where the player can get drunk and where regular customers hang out, talk, argue, or react to the player.\n\nThe game should include classic GTA-style missions, but they should not feel repetitive. Each mission should be slightly different. Add around 20–30 missions in total. Mission types can include:\n\n* delivering cargo from point A to point B,\n* assassinating a gang leader,\n* stealing a car full of money,\n* winning illegal street races,\n* escaping police during a race,\n* finding and eliminating a rat cooperating with the police,\n* protecting a character,\n* attacking a gang hideout,\n* stealing weapons,\n* escaping by boat through canals,\n* chasing a target through traffic,\n* and other open-world crime-action mission types.\n\nThe game should include a mission system with objectives, markers, failure conditions, success conditions, rewards, and mission progression. Missions should use different vehicles, locations, enemies, and situations so that they do not feel copied from each other.\n\nAdd sounds for everything important:\n\n* car engines,\n* horns,\n* brakes,\n* crashes,\n* tires,\n* boats,\n* helicopters,\n* airplanes,\n* footsteps,\n* punches,\n* gunfire,\n* explosions,\n* people talking,\n* frightened civilians screaming,\n* angry civilians shouting,\n* police sirens,\n* gang threats,\n* bar ambience,\n* beach ambience,\n* city traffic,\n* ocean waves,\n* and general city noise.\n\nBecause this is a large and demanding task, plan it carefully before implementation. First create a dedicated TODO.md file. In TODO.md, break the whole project into clear milestones and smaller tasks. Then complete the work step by step, following the TODO.md file point by point. Keep updating TODO.md as tasks are completed. Do not skip details. Build the game gradually, starting with the core playable systems, then the city, traffic, pedestrians, police, gangs, missions, sound, UI, and final polish."
  },
  "provenancePolicy": {
    "sk": "Dátum, verzia promptu a typ behu sú uvedené iba vtedy, keď boli zaznamenané; historické medzery sa neodhadujú.",
    "en": "A run date, prompt version, and run type are included only when recorded; historical gaps are not estimated."
  },
  "runs": [
    {
      "rank": 1,
      "model": "Claude Fable 5",
      "publisher": "Anthropic",
      "modelId": null,
      "runtime": {
        "sk": "Uzavretý",
        "en": "Closed"
      },
      "parameters": null,
      "requirements": null,
      "evaluation": {
        "sk": {
          "assignment": "Splnil zadanie",
          "playability": "Hrateľný",
          "bugs": "Veľa chýb",
          "note": "1 opravný prompt. Generovanie približne 1 h 20 min; takmer 400K kontext."
        },
        "en": {
          "assignment": "Matched the assignment",
          "playability": "Playable",
          "bugs": "Many bugs",
          "note": "1 corrective prompt. About 1 h 20 min generation; nearly 400K context."
        }
      },
      "description": {
        "sk": "Najvyššie hodnotený hrateľný výsledok GTA benchmarku. Rozsiahly svet funguje, ale stále obsahuje veľa chýb.",
        "en": "The top-ranked playable GTA benchmark result. Its extensive world works, but it still contains many bugs."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/claude-fable5.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/fable-5/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": null
    },
    {
      "rank": 2,
      "model": "Claude Opus 5",
      "publisher": "Anthropic",
      "modelId": null,
      "runtime": {
        "sk": "Uzavretý",
        "en": "Closed"
      },
      "parameters": null,
      "requirements": null,
      "evaluation": {
        "sk": {
          "assignment": "Splnil zadanie čiastočne",
          "playability": "Hrateľný",
          "bugs": "Veľmi veľa chýb",
          "note": "1 opravný prompt na ovládanie."
        },
        "en": {
          "assignment": "Partially matched",
          "playability": "Playable",
          "bugs": "Very many bugs",
          "note": "1 corrective prompt for controls."
        }
      },
      "description": {
        "sk": "Rozsiahly open-world prototyp Neon Bay je hrateľný, ale obsahuje veľmi veľké množstvo vážnych chýb.",
        "en": "The extensive Neon Bay open-world prototype is playable, but contains a very large number of serious bugs."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/claude-opus5.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/claude-opus-5/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": null
    },
    {
      "rank": 3,
      "model": "GPT-5.6 Sol",
      "publisher": "OpenAI",
      "modelId": null,
      "runtime": {
        "sk": "Uzavretý",
        "en": "Closed"
      },
      "parameters": null,
      "requirements": null,
      "evaluation": {
        "sk": {
          "assignment": "Splnil zadanie čiastočne",
          "playability": "Hrateľný",
          "bugs": "Veľmi veľa chýb",
          "note": "Vyžadoval 2 opravné prompty na ovládanie."
        },
        "en": {
          "assignment": "Partially matched",
          "playability": "Playable",
          "bugs": "Very many bugs",
          "note": "Required 2 corrective prompts for controls."
        }
      },
      "description": {
        "sk": "Ambiciózny open-world prototyp Neon Palms je hrateľný, ale obsahuje veľmi veľa chýb.",
        "en": "The ambitious Neon Palms open-world prototype is playable, but contains very many bugs."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/gpt5-6-sol.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/gpt-5-6-sol/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": null
    },
    {
      "rank": 4,
      "model": "GPT-5.5",
      "publisher": "OpenAI",
      "modelId": null,
      "runtime": {
        "sk": "Uzavretý",
        "en": "Closed"
      },
      "parameters": null,
      "requirements": null,
      "evaluation": {
        "sk": {
          "assignment": "Splnil zadanie čiastočne",
          "playability": "Hrateľný",
          "bugs": "Veľmi veľa chýb",
          "note": ""
        },
        "en": {
          "assignment": "Partially matched",
          "playability": "Playable",
          "bugs": "Very many bugs",
          "note": ""
        }
      },
      "description": {
        "sk": "Vizuálne rozpoznateľný open-world prototyp je hrateľný, ale obsahuje veľmi veľa chýb.",
        "en": "A visually recognizable open-world prototype that is playable, but contains very many bugs."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/gpt5-5.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/gpt-5-5/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": null
    },
    {
      "rank": 5,
      "model": "Gemini 3.5 Flash",
      "publisher": "Alphabet",
      "modelId": null,
      "runtime": {
        "sk": "Uzavretý",
        "en": "Closed"
      },
      "parameters": null,
      "requirements": null,
      "evaluation": {
        "sk": {
          "assignment": "Splnil zadanie čiastočne",
          "playability": "Nehrateľný",
          "bugs": "Veľmi veľa chýb",
          "note": ""
        },
        "en": {
          "assignment": "Partially matched",
          "playability": "Unplayable",
          "bugs": "Very many bugs",
          "note": ""
        }
      },
      "description": {
        "sk": "Ambiciózny, ale pre veľmi veľa chýb nehrateľný open-world výstup.",
        "en": "An ambitious open-world output that is unplayable because of very many bugs."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/gemini-3-5-flash.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/gemini-3-5-flash/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": null
    },
    {
      "rank": 6,
      "model": "GLM 5.2",
      "publisher": "Zhipu AI",
      "modelId": "glm-5-2",
      "runtime": {
        "sk": "Otvorené váhy (OpenRouter)",
        "en": "Open-weight (OpenRouter)"
      },
      "parameters": {
        "sk": "744B",
        "en": "744B"
      },
      "requirements": {
        "sk": [
          {
            "kind": "weights",
            "text": "464,8 GB"
          },
          {
            "kind": "context",
            "text": "≈9,52 GB / 100K"
          }
        ],
        "en": [
          {
            "kind": "weights",
            "text": "464.8 GB"
          },
          {
            "kind": "context",
            "text": "≈9.52 GB / 100K"
          }
        ]
      },
      "evaluation": {
        "sk": {
          "assignment": "Splnil zadanie čiastočne",
          "playability": "Nehrateľný",
          "bugs": "Veľmi veľa chýb",
          "note": ""
        },
        "en": {
          "assignment": "Partially matched",
          "playability": "Unplayable",
          "bugs": "Very many bugs",
          "note": ""
        }
      },
      "description": {
        "sk": "Open-world výstup s veľmi veľa chybami, ktoré znemožňujú praktické hranie.",
        "en": "An open-world output with very many bugs that prevent practical play."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/glm-5-2.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/glm-5-2/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": {
        "modelRevision": "zai-org/GLM-5.2@b4734de4facf877f85769a911abafc5283eab3d9",
        "checkedAt": "2026-08-07",
        "sources": [
          {
            "id": "glm-5-2-model",
            "kind": "official checkpoint",
            "url": "https://huggingface.co/zai-org/GLM-5.2/tree/b4734de4facf877f85769a911abafc5283eab3d9",
            "revision": "b4734de4facf877f85769a911abafc5283eab3d9",
            "accessedAt": "2026-08-07"
          },
          {
            "id": "glm-5-2-nvfp4",
            "kind": "vendor 4-bit conversion",
            "url": "https://huggingface.co/nvidia/GLM-5.2-NVFP4/tree/aec724e8c7b8ee9db3b48c01c320f63f9cdaf8aa",
            "revision": "aec724e8c7b8ee9db3b48c01c320f63f9cdaf8aa",
            "accessedAt": "2026-08-07"
          },
          {
            "id": "vllm-cache-layout",
            "kind": "runtime source",
            "url": "https://github.com/vllm-project/vllm/tree/946452961265514622b95ea170839b214f39c2a0",
            "revision": "946452961265514622b95ea170839b214f39c2a0",
            "accessedAt": "2026-08-07"
          }
        ]
      }
    },
    {
      "rank": 7,
      "model": "Qwen 3.6 27B ThinkingCap",
      "publisher": "Alibaba",
      "modelId": "thinkingcap-qwen3-6-27b",
      "runtime": {
        "sk": "Otvorené váhy",
        "en": "Open-weight"
      },
      "parameters": {
        "sk": "27B",
        "en": "27B"
      },
      "requirements": {
        "sk": [
          {
            "kind": "weights",
            "text": "16,8 GB"
          },
          {
            "kind": "context",
            "text": "≈6,71 GB / 100K"
          }
        ],
        "en": [
          {
            "kind": "weights",
            "text": "16.8 GB"
          },
          {
            "kind": "context",
            "text": "≈6.71 GB / 100K"
          }
        ]
      },
      "evaluation": {
        "sk": {
          "assignment": "Zadanie nesplnil",
          "playability": "Nefungoval",
          "bugs": "Obsahuje fatálne chyby",
          "note": "Žiaden viditeľný svet.\nGenerovanie približne 1 h 17 min."
        },
        "en": {
          "assignment": "Did not match the assignment",
          "playability": "Did not work",
          "bugs": "Contains fatal bugs",
          "note": "No visible world.\nAbout 1 h 17 min generation."
        }
      },
      "description": {
        "sk": "Zadanie nesplnil, nefungoval a neobsahuje žiaden viditeľný svet.",
        "en": "It did not match the assignment, did not work, and contains no visible world."
      },
      "screenshot": "https://dubovskylabs.com/assets/img/benchmarks/gta/qwen3-6-27b-thinkingcap.png",
      "demo": "https://dubovskylabs.com/benchmarks/gta/play/qwen-3-6-27b-thinkingcap/index.html",
      "provenance": {
        "generatedAt": null,
        "testedAt": null,
        "promptVersion": null,
        "runType": null
      },
      "modelSource": {
        "modelRevision": "bottlecapai/ThinkingCap-Qwen3.6-27B-GGUF@0982db1be5e1e6cf7163ba89d7f63a9f18d2f4f0",
        "checkedAt": "2026-08-07",
        "sources": [
          {
            "id": "thinkingcap-gguf",
            "kind": "publisher checkpoint",
            "url": "https://huggingface.co/bottlecapai/ThinkingCap-Qwen3.6-27B-GGUF/tree/0982db1be5e1e6cf7163ba89d7f63a9f18d2f4f0",
            "revision": "0982db1be5e1e6cf7163ba89d7f63a9f18d2f4f0",
            "accessedAt": "2026-08-07"
          },
          {
            "id": "qwen3-6-27b-config",
            "kind": "official config",
            "url": "https://huggingface.co/Qwen/Qwen3.6-27B/blob/6a9e13bd6fc8f0983b9b99948120bc37f49c13e9/config.json",
            "revision": "6a9e13bd6fc8f0983b9b99948120bc37f49c13e9",
            "accessedAt": "2026-08-07"
          },
          {
            "id": "transformers-cache",
            "kind": "runtime source",
            "url": "https://github.com/huggingface/transformers/tree/42f189ded85d18d00b51161d694cafd325e32b91",
            "revision": "42f189ded85d18d00b51161d694cafd325e32b91",
            "accessedAt": "2026-08-07"
          }
        ]
      }
    }
  ]
}