{"slug": "godot-benchmark-sol-k3-fable", "title": "Godot Benchmark: Sol > K3 > Fable", "summary": "A benchmark of three AI models for game development found that GPT 5.6 Sol outperforms Kimi 3 and Claude Fable 5 in speed and cost, despite producing slightly worse results. Sol completed a 3D parkour world in 9.7 minutes at $2.94, while Kimi took 79 minutes at $3.14 and Fable took 32.5 minutes at $13. The test, which involved creating a 3D parkour world with three races using Kenney assets, rated Sol's conversation efficiency at 9/10, Kimi's thoroughness at 7/10, and Fable's debugging at 6/10.", "body_md": "# Godot Benchmark: Sol > K3 > Fable\n\nSo here’s the thing - every model is benchmaxxing, so how do you know which one is *actually* the best to use? The thing we’ve found most reliable is to benchmark it with something that’s relevant to you.\n\n## TL;DR\n\nKimi actually beat Fable, but it’s unusably slow. Sol’s speed and affordability make it better for game development, even if the end result is slightly worse than the other two.\n\n| Model | World | Conversation | Cost | Time |\n|---|---|---|---|---|\nGPT 5.6 Sol | 6 | 9 | $2.94 | 9.7 min |\nKimi 3 | 8 | 7 | $3.14 | 79 min |\nClaude Fable 5 | 7 | 6 | $13* | 32.5 min |\n\n## The Prompt\n\nMake a 3D parkour world where you compete against bots for time. There should be 3 races in the world, and you can start any of the races by walking over to their start point. Use the assets in Kenney_Platformer_Kit/ to build your world.\n\n## Created Worlds\n\n| GPT 5.6 Sol | Kimi 3 | Claude Fable 5 |\n|---|---|---|\n|\n\n[Play](https://play.ziva.sh/buttery-earnestness/)[Play](https://play.ziva.sh/hypnotized-snail/)## Conversation Analysis\n\nI ran each conversation transcript through Opus 4.8 and had it come up with a grade:\n\n**GPT 5.6 Sol**— tightest loop; headless logic checks, playtests only where they counted.** 9/10**·[conversation](https://ziva.sh/c/mrAGKPVAFY)** Kimi 3**— most thorough of the three (12 playtests), but redundant and slow.** 7/10**·[conversation](https://ziva.sh/c/DaApZ2hVrb)** Claude Fable 5**— sharpest debugging, but ships blind: it never sees its own playtest frames.** 6/10**·[conversation](https://ziva.sh/c/S8UvSEhnB_)\n\n## My Thoughts\n\nKimi is lowkey very good here, but it’s painfully slow. Maybe it’s a reliability issue on their GPUs and it’ll be better once more people can self-host, but the truth is right now the time it takes to do things makes it borderline unusable.\n\n## Credits\n\nAssets used for the game generation are provided by Kenny.nl", "url": "https://wpnews.pro/news/godot-benchmark-sol-k3-fable", "canonical_source": "https://ziva.sh/blogs/godot-ai-benchmark", "published_at": "2026-07-19 00:00:00+00:00", "updated_at": "2026-07-27 00:35:42.411958+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-tools", "ai-products", "generative-ai", "ai-research"], "entities": ["GPT 5.6 Sol", "Kimi 3", "Claude Fable 5", "Kenney", "Opus 4.8", "Ziva"], "alternates": {"html": "https://wpnews.pro/news/godot-benchmark-sol-k3-fable", "markdown": "https://wpnews.pro/news/godot-benchmark-sol-k3-fable.md", "text": "https://wpnews.pro/news/godot-benchmark-sol-k3-fable.txt", "jsonld": "https://wpnews.pro/news/godot-benchmark-sol-k3-fable.jsonld"}}