Best AI model for game development
In game development, work within 60 Elo of the best model is available for 2× less: Claude Sonnet 5.5 at $4.03 per thousand tasks against $8.86 for Claude Opus 5.5.
112 models rated head to head in this category, not scored by a general benchmark and assumed to transfer. Playable mechanics built from a prompt: movement, collision, state and win conditions, compared head to head.
Highest ratedClaude Opus 5.51467 Elo · $8.86 / 1k tasks
Best valueClaude Sonnet 5.51458 Elo · $4.03 / 1k tasks
Nothing cheaper is as good
8 models are undominated here: for each, no cheaper model scores within 18 Elo. Everything else in the table below is beaten on both counts at once.
- 1Ministral 3 3B 2512954 Elo$0.085 / 1k131K context
- 2Solar Mini 41136 Elo$0.088 / 1k524K context
- 3Solar Pro 41176 Elo$0.164 / 1k524K context
- 4DeepSeek V4 Flash 04231202 Elo$0.170 / 1k1.05M contextfp4 on some hosts
- 5MiMo-V2.51250 Elo$0.183 / 1k1.05M context
- 6Claude Haiku 5.51307 Elo$0.215 / 1k1.00M context
- 7Muse Spark 1.31351 Elo$1.98 / 1k1.05M context
- 8Claude Sonnet 5.51458 Elo$4.03 / 1k1.00M context
Every model rated on game development
Elo measured in this category specifically. A model strong overall can sit well down this table, which is the whole reason for ranking by task.
| # | Model | Elo | $/1k | Context | Hosts |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | 1467 | $8.86 | 1.00M | 11 |
| 2 | Claude Sonnet 5.5 | 1458 | $4.03 | 1.00M | 8 |
| 3 | Claude Fable 5.1 | 1387 | $19.98 | 1.00M | 4 |
| 4 | GPT-6.1 Sol | 1375 | $4.43 | 1.05M | 8 |
| 5 | Kimi K3 | 1369 | $6.24 | 1.05M | 25 |
| 6 | Muse Spark 1.3 | 1351 | $1.98 | 1.05M | 1 |
| 7 | Claude Opus 5 | 1343 | $11.25 | 1.00M | 11 |
| 8 | Claude Fable 5 | 1341 | $20.52 | 1.00M | 6 |
| 9 | Gemini 3.7 Flash | 1326 | $1.54 | 1.05M | 6 |
| 10 | GLM 5.3 | 1319 | $1.93 | 1.05M | 41 |
| 11 | MiMo-V2.6-Pro | 1318 | $0.472 | 1.05M | 4 |
| 12 | Gemini 3.8 Flash | 1312 | $1.54 | 1.05M | 6 |
| 13 | Muse Spark 1.2 | 1308 | $1.98 | 1.05M | 1 |
| 14 | Claude Haiku 5.5 | 1307 | $0.215 | 1.00M | 10 |
| 15 | Grok 4.7 | 1305 | $3.40 | 500K | 5 |
| 16 | Grok 4.6 | 1300 | $3.40 | 500K | 6 |
| 17 | GPT-5.5 | 1299 | $12.90 | 1.05M | 7 |
| 18 | GLM 5.3 Flash | 1287 | $0.230 | 1.05M | 32 |
| 19 | Claude Sonnet 5 | 1287 | $4.30 | 1.00M | 10 |
| 20 | Claude Opus 4.6 | 1283 | $10.26 | 1.00M | 6 |
| 21 | Muse Spark 1.1 | 1276 | $1.98 | 1.05M | 1 |
| 22 | GLM 5.2 | 1273 | $2.00 | 1.05M | 34 |
| 23 | Qwen3.7 Max | 1272 | $2.61 | 1.00M | 2 |
| 24 | Gemini 3.6 Flash | 1268 | $1.54 | 1.05M | 7 |
| 25 | Claude Opus 4.8 | 1268 | $11.25 | 1.00M | 11 |
| 26 | Gemini 3.5 Flash | 1268 | $3.53 | 1.05M | 7 |
| 27 | GLM 5.1 | 1268 | $2.11 | 203K | 13 |
| 28 | Grok 4.5 | 1267 | $4.36 | 500K | 4 |
| 29 | MiMo-V2.5-Pro | 1265 | $0.497 | 1.05M | 6 |
| 30 | Qwen3.7 Plus | 1264 | $0.584 | 1.00M | 2 |
| 31 | Claude Sonnet 4.6 | 1264 | $6.16 | 1.00M | 9 |
| 32 | Kimi K2.6 | 1256 | $1.52 | 262K | 17 |
| 33 | GLM 5 Turbo | 1254 | $1.95 | 203K | 1 |
| 34 | MiMo-V2.5 | 1250 | $0.183 | 1.05M | 5 |
| 35 | GPT-5.4 | 1244 | $6.45 | 1.05M | 7 |
| 36 | GLM 5 | 1239 | $1.42 | 203K | 8 |
| 37 | DeepSeek V4 Pro 0423 | 1234 | $1.68 | 1.05M | 15 |
| 38 | Claude Opus 4.5 | 1234 | $10.26 | 200K | 6 |
| 39 | Mistral Large 4 | 1225 | $1.00 | 1.05M | 3 |
| 40 | Kimi K2.7 Code | 1224 | $1.75 | 262K | 12 |
| 41 | GLM 5V Turbo | 1223 | $1.95 | 203K | 1 |
| 42 | MiniMax M3 | 1218 | $0.547 | 1.05M | 13 |
| 43 | Qwen3.6 Plus | 1217 | $0.975 | 1.00M | 1 |
| 44 | MiniMax M2.7 | 1216 | $0.534 | 197K | 6 |
| 45 | DeepSeek V4 Flash 0731 | 1214 | $0.233 | 1.05M | 27 |
| 46 | Kimi K2.5 | 1212 | $1.22 | 262K | 5 |
| 47 | Gemini 3.1 Pro Preview | 1208 | $4.70 | 1.05M | 6 |
| 48 | DeepSeek V4 Flash 0423 | 1202 | $0.170 | 1.05M | 16 |
| 49 | GPT-5.2 | 1200 | $5.17 | 400K | 4 |
| 50 | Grok 4.20 | 1198 | $2.17 | 2.00M | 4 |
| 51 | GLM 4.7 | 1194 | $0.980 | 203K | 6 |
| 52 | GPT-5 | 1190 | $3.69 | 400K | 3 |
| 53 | GPT-5.1 | 1184 | $3.69 | 400K | 5 |
| 54 | MiniMax M2.5 | 1182 | $0.523 | 197K | 9 |
| 55 | Grok 4.3 | 1180 | $2.17 | 1.00M | 4 |
| 56 | Claude Opus 4.1 | 1177 | $30.78 | 200K | 1 |
| 57 | Solar Pro 4 | 1176 | $0.164 | 524K | 2 |
| 58 | Gemini 3 Flash Preview | 1171 | $1.18 | 1.05M | 6 |
| 59 | Claude Sonnet 4.5 | 1171 | $6.16 | 1.00M | 7 |
| 60 | Inkling | 1169 | $1.81 | 524K | 2 |
| 61 | GPT-5.3-Codex | 1167 | $5.17 | 400K | 3 |
| 62 | Step 3.7 Flash | 1165 | $0.470 | 262K | 2 |
| 63 | GLM 4.6 | 1156 | $0.954 | 203K | 4 |
| 64 | DeepSeek V3.2 Exp | 1150 | $0.447 | 164K | 1 |
| 65 | GLM 4.5 | 1150 | $1.03 | 131K | 1 |
| 66 | Qwen3.5 397B A17B | 1146 | $1.49 | 262K | 10 |
| 67 | Claude Sonnet 4 | 1144 | $6.16 | 200K | 2 |
| 68 | GPT-5.1-Codex | 1143 | $3.69 | 400K | 1 |
| 69 | Nemotron 3 Ultra | 1142 | $1.09 | 203K | 4 |
| 70 | Hy3 | 1141 | $0.265 | 262K | 5 |
| 71 | GLM 4.7 Flash | 1140 | $0.193 | 128K | 3 |
| 72 | MiniMax M2.1 | 1140 | $0.706 | 205K | 2 |
| 73 | DeepSeek V3.2 | 1137 | $0.504 | 164K | 12 |
| 74 | Solar Mini 4 | 1136 | $0.088 | 524K | 2 |
| 75 | DeepSeek V3.1 Terminus | 1136 | $0.672 | 164K | 2 |
| 76 | GPT-5 Mini | 1135 | $0.738 | 400K | 4 |
| 77 | MiniMax M2 | 1123 | $0.666 | 205K | 2 |
| 78 | Gemini 2.5 Pro | 1117 | $3.69 | 1.05M | 7 |
| 79 | Qwen3.5 Plus 2026-02-15 | 1111 | $0.780 | 1.00M | 1 |
| 80 | Qwen3 Coder 480B A35B | 1104 | $0.666 | 262K | 3 |
| 81 | R1 0528 | 1103 | $1.18 | 164K | 4 |
| 82 | Claude Haiku 4.5 | 1102 | $2.15 | 200K | 8 |
| 83 | GLM 4.5 Air | 1099 | $0.343 | 131K | 3 |
| 84 | GPT-5.1-Codex-Mini | 1098 | $0.742 | 400K | 1 |
| 85 | DeepSeek V3.1 | 1089 | $0.808 | 164K | 5 |
| 86 | Mistral Large 3 2512 | 1085 | $0.760 | 262K | 2 |
| 87 | GPT-4.1 | 1085 | $3.72 | 1.05M | 3 |
| 88 | Trinity Large Thinking | 1084 | $0.403 | 262K | 1 |
| 89 | Mistral Medium 3.1 | 1079 | $0.821 | 131K | 3 |
| 90 | GPT-4.1 Mini | 1076 | $0.744 | 1.05M | 3 |
| 91 | Gemini 2.5 Flash | 1071 | $0.916 | 1.05M | 7 |
| 92 | DeepSeek V3 | 1059 | $0.634 | 128K | 2 |
| 93 | GPT-5 Nano | 1050 | $0.148 | 400K | 4 |
| 94 | Ministral 3 14B 2512 | 1043 | $0.170 | 262K | 3 |
| 95 | o3 | 1040 | $3.72 | 200K | 1 |
| 96 | Gemini 3.1 Flash Lite Preview | 1034 | $0.588 | 1.05M | 3 |
| 97 | Mistral Medium 3 | 1023 | $0.821 | 131K | 3 |
| 98 | o4 Mini | 1008 | $2.05 | 200K | 1 |
| 99 | gpt-oss-120b | 997 | $0.285 | 131K | 24 |
| 100 | Ministral 3 8B 2512 | 996 | $0.128 | 262K | 3 |
| 101 | Mercury 2 | 991 | $0.363 | 128K | 1 |
| 102 | Kimi K2 0711 | 979 | $1.37 | 131K | 1 |
| 103 | GPT-4.1 Nano | 976 | $0.186 | 1.05M | 3 |
| 104 | Codestral 2508 | 973 | $0.456 | 256K | 2 |
| 105 | Qwen3 235B A22B Thinking 2507 | 962 | $1.59 | 128K | 1 |
| 106 | Qwen3 235B A22B Instruct 2507 | 957 | $0.405 | 262K | 7 |
| 107 | Ministral 3 3B 2512 | 954 | $0.085 | 131K | 3 |
| 108 | GPT-4o | 911 | $5.10 | 128K | 2 |
| 109 | Qwen3 30B A3B | 904 | $0.294 | 41K | 1 |
| 110 | Mistral Small 3.2 24B | 894 | $0.193 | 128K | 4 |
| 111 | Llama 4 Maverick | 843 | $0.462 | 524K | 4 |
| 112 | Llama 4 Scout | 778 | $0.393 | 328K | 3 |
Other tasks
Websitessave 4×Code generationsave 4×Data visualisationsave 4×UI componentssave 19×3D graphicssave 2×SVG and vectorsave 39×ASCII artsave 6×
Elo comes from head-to-head build-offs via OpenRouter, compiled 12 October 2026 and refreshed daily. Cost is the median across every host serving each model, priced on 1,200 input and 300 output tokens with 60% cached. See the findings for method, or download the data.