Best AI model for SVG and vector
In SVG and vector, work within 60 Elo of the best model is available for 39× less: GLM 5.3 Flash at $0.230 per thousand tasks against $8.86 for Claude Opus 5.5.
79 models rated head to head in this category, not scored by a general benchmark and assumed to transfer. Vector output written directly as markup for icons, diagrams and illustrations, where a model must reason about coordinates it cannot see.
Nothing cheaper is as good
4 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.
- 1DeepSeek V4 Flash 04231155 Elo$0.170 / 1k1.05M contextfp4 on some hosts
- 2MiMo-V2.51176 Elo$0.183 / 1k1.05M context
- 3GLM 5.3 Flash1289 Elo$0.230 / 1k1.05M contextfp4 on some hosts
- 4Muse Spark 1.31324 Elo$1.98 / 1k1.05M context
Every model rated on SVG and vector
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 | 1334 | $8.86 | 1.00M | 11 |
| 2 | Muse Spark 1.3 | 1324 | $1.98 | 1.05M | 1 |
| 3 | Claude Opus 5 | 1314 | $11.25 | 1.00M | 11 |
| 4 | Kimi K3 | 1305 | $6.24 | 1.05M | 25 |
| 5 | Claude Fable 5 | 1303 | $20.52 | 1.00M | 6 |
| 6 | Claude Fable 5.1 | 1302 | $19.98 | 1.00M | 4 |
| 7 | GLM 5.3 | 1294 | $1.93 | 1.05M | 41 |
| 8 | Gemini 3.8 Flash | 1293 | $1.54 | 1.05M | 6 |
| 9 | GLM 5.3 Flash | 1289 | $0.230 | 1.05M | 32 |
| 10 | Muse Spark 1.2 | 1289 | $1.98 | 1.05M | 1 |
| 11 | Gemini 3.1 Pro Preview | 1285 | $4.70 | 1.05M | 6 |
| 12 | Gemini 3.7 Flash | 1280 | $1.54 | 1.05M | 6 |
| 13 | Gemini 3.5 Flash | 1253 | $3.53 | 1.05M | 7 |
| 14 | Grok 4.6 | 1241 | $3.40 | 500K | 6 |
| 15 | Muse Spark 1.1 | 1236 | $1.98 | 1.05M | 1 |
| 16 | GPT-5.5 | 1235 | $12.90 | 1.05M | 7 |
| 17 | Mistral Large 4 | 1232 | $1.00 | 1.05M | 3 |
| 18 | Claude Opus 4.6 | 1227 | $10.26 | 1.00M | 6 |
| 19 | GLM 5.1 | 1218 | $2.11 | 203K | 13 |
| 20 | Grok 4.5 | 1217 | $4.36 | 500K | 4 |
| 21 | GLM 5.2 | 1217 | $2.00 | 1.05M | 34 |
| 22 | Qwen3.7 Max | 1216 | $2.61 | 1.00M | 2 |
| 23 | Qwen3.7 Plus | 1207 | $0.584 | 1.00M | 2 |
| 24 | GLM 5 Turbo | 1206 | $1.95 | 203K | 1 |
| 25 | Claude Sonnet 4.6 | 1196 | $6.16 | 1.00M | 9 |
| 26 | GPT-5.4 | 1194 | $6.45 | 1.05M | 7 |
| 27 | Claude Opus 4.8 | 1193 | $11.25 | 1.00M | 11 |
| 28 | Claude Sonnet 5 | 1192 | $4.30 | 1.00M | 10 |
| 29 | GPT-5 | 1188 | $3.69 | 400K | 3 |
| 30 | DeepSeek V4 Flash 0731 | 1186 | $0.233 | 1.05M | 27 |
| 31 | Kimi K2.6 | 1181 | $1.52 | 262K | 17 |
| 32 | Claude Opus 4.5 | 1180 | $10.26 | 200K | 6 |
| 33 | MiMo-V2.5-Pro | 1177 | $0.497 | 1.05M | 6 |
| 34 | Kimi K2.7 Code | 1176 | $1.75 | 262K | 12 |
| 35 | MiMo-V2.5 | 1176 | $0.183 | 1.05M | 5 |
| 36 | GLM 5 | 1167 | $1.42 | 203K | 8 |
| 37 | Grok 4.20 | 1164 | $2.17 | 2.00M | 4 |
| 38 | MiniMax M3 | 1162 | $0.547 | 1.05M | 13 |
| 39 | Qwen3.6 Plus | 1161 | $0.975 | 1.00M | 1 |
| 40 | DeepSeek V4 Flash 0423 | 1155 | $0.170 | 1.05M | 16 |
| 41 | GLM 5V Turbo | 1151 | $1.95 | 203K | 1 |
| 42 | Claude Opus 4.1 | 1151 | $30.78 | 200K | 1 |
| 43 | Kimi K2.5 | 1149 | $1.22 | 262K | 5 |
| 44 | MiniMax M2.5 | 1148 | $0.523 | 197K | 9 |
| 45 | GPT-5.1 | 1148 | $3.69 | 400K | 5 |
| 46 | GLM 4.7 | 1142 | $0.980 | 203K | 6 |
| 47 | DeepSeek V4 Pro 0423 | 1139 | $1.68 | 1.05M | 15 |
| 48 | Qwen3.5 397B A17B | 1139 | $1.49 | 262K | 10 |
| 49 | GPT-5.2 | 1138 | $5.17 | 400K | 4 |
| 50 | MiniMax M2.7 | 1136 | $0.534 | 197K | 6 |
| 51 | MiniMax M2.1 | 1131 | $0.706 | 205K | 2 |
| 52 | GPT-5.3-Codex | 1128 | $5.17 | 400K | 3 |
| 53 | Claude Sonnet 4.5 | 1114 | $6.16 | 1.00M | 7 |
| 54 | Inkling | 1112 | $1.81 | 524K | 2 |
| 55 | GLM 4.6 | 1108 | $0.954 | 203K | 4 |
| 56 | Qwen3.5 Plus 2026-02-15 | 1107 | $0.780 | 1.00M | 1 |
| 57 | MiniMax M2 | 1099 | $0.666 | 205K | 2 |
| 58 | GLM 4.5 | 1095 | $1.03 | 131K | 1 |
| 59 | GPT-5 Mini | 1090 | $0.738 | 400K | 4 |
| 60 | Step 3.7 Flash | 1082 | $0.470 | 262K | 2 |
| 61 | Grok 4.3 | 1082 | $2.17 | 1.00M | 4 |
| 62 | Claude Sonnet 4 | 1077 | $6.16 | 200K | 2 |
| 63 | GLM 4.5 Air | 1070 | $0.343 | 131K | 3 |
| 64 | Nemotron 3 Ultra | 1068 | $1.09 | 203K | 4 |
| 65 | DeepSeek V3.1 Terminus | 1064 | $0.672 | 164K | 2 |
| 66 | Gemini 3.1 Flash Lite Preview | 1051 | $0.588 | 1.05M | 3 |
| 67 | R1 0528 | 1038 | $1.18 | 164K | 4 |
| 68 | GLM 4.7 Flash | 1037 | $0.193 | 128K | 3 |
| 69 | DeepSeek V3.2 | 1031 | $0.504 | 164K | 12 |
| 70 | DeepSeek V3.2 Exp | 1031 | $0.447 | 164K | 1 |
| 71 | Claude Haiku 4.5 | 1025 | $2.15 | 200K | 8 |
| 72 | Gemini 2.5 Flash | 1019 | $0.916 | 1.05M | 7 |
| 73 | Trinity Large Thinking | 1015 | $0.403 | 262K | 1 |
| 74 | Mercury 2 | 997 | $0.363 | 128K | 1 |
| 75 | Mistral Large 3 2512 | 992 | $0.760 | 262K | 2 |
| 76 | Mistral Medium 3.1 | 992 | $0.821 | 131K | 3 |
| 77 | GPT-5.1-Codex-Mini | 977 | $0.742 | 400K | 1 |
| 78 | DeepSeek V3 | 975 | $0.634 | 128K | 2 |
| 79 | DeepSeek V3.1 | 966 | $0.808 | 164K | 5 |
Other tasks
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.