Coding Models
Updated
Which model writes the better working software, from real engineering tests and developers' side-by-side votes.
Coding model leaderboard
151 models · ToolsGuide Score out of 1,000, where 1,000 is a flawless model
20 of 151 models in the chart · ·
All models
| Compare | # | Model | Creator | Licence | Input / 1M | Score | From votes | From tests · count | Votes |
|---|---|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 (Max) |   Anthropic | Proprietary | $4.00 | 648 | 675 | 630 · 6 | 3,473 | |
| 2 | Gemini 4 Carbon (High) |  Google | Proprietary | – | 628 | 628 | – | 2,230 | |
| 3 | Claude Sonnet 5.5 (xHigh) | Anthropic | Proprietary | $2.00 | 609 | 630 | 595 · 6 | 4,469 | |
| 4 | Claude Fable 5.1 (Max) | Anthropic | Proprietary | $10.00 | 603 | 603 | – | 8,614 | |
| 5 | GPT-6 Astra (Max) |   OpenAI | Proprietary | $10.00 | 589 | 634 | 558 · 3 | 7,210 | |
| 6 | Claude Sonnet 5.5 (High) | Anthropic | Proprietary | $2.00 | 582 | 563 | 595 · 6 | 6,371 | |
| 7 | Kimi K3.1 Pro (Max) |   Moonshot AI | Proprietary | $2.00 | 540 | 545 | 533 · 2 | 2,671 | |
| 8 | Gemini 4 Argon (High) | Google | Proprietary | $2.00 | 539 | 528 | 556 · 2 | 3,437 | |
| 9 | Claude Opus 5 (Max) | Anthropic | Proprietary | $5.00 | 538 | 540 | 537 · 3 | 19,641 | |
| 10 | Claude Opus 5 (High) | Anthropic | Proprietary | $5.00 | 524 | 505 | 537 · 3 | 26,887 | |
| 11 | Muse Wonder (Max) |  Meta | Proprietary | $4.00 | 524 | 499 | 541 · 3 | 3,724 | |
| 12 | GPT-6.1 Terra (Max) | OpenAI | Proprietary | $1.00 | 521 | 521 | – | 1,821 | |
| 13 | GPT-6.1 Sol (Max) | OpenAI | Proprietary | $2.00 | 520 | 608 | 405 · 2 | 4,092 | |
| 14 | DeepSeek-V4.1-Flash-20261017 (Max) |  DeepSeek | MIT | – | 513 | 513 | – | 7,842 | |
| 15 | Muse Wonder Full (Max) | Meta | Proprietary | – | 509 | 579 | 415 · 2 | 2,248 | |
| 16 | Grok 4.7 (xHigh) |   xAI | Proprietary | $2.00 | 488 | 484 | 494 · 2 | 6,572 | |
| 17 | Claude Haiku 5.5 (High) | Anthropic | Proprietary | $0.10 | 474 | 436 | 501 · 5 | 4,758 | |
| 18 | MiniMax-M3.1-Flash |  MiniMax | MiniMax Community License | $0.15 | 471 | 430 | 501 · 3 | 3,062 | |
| 19 | GPT-6 Sol (Max) | OpenAI | Proprietary | $2.00 | 467 | 527 | 387 · 2 | 5,714 | |
| 20 | DeepSeek-V4.1-Flash-20260910 (Max) | DeepSeek | MIT | – | 466 | 466 | – | 3,510 | |
| 21 | GPT-5.6 Sol (xHigh, Codex Harness) | OpenAI | Proprietary | $5.00 | 465 | 462 | 467 · 6 | 18,884 | |
| 22 | MiMo-V2.6-Pro |  Xiaomi | MIT | $0.43 | 461 | 472 | 455 · 4 | 6,820 | |
| 23 | Mistral Large 4 |  Mistral | Proprietary | $0.70 | 460 | 396 | 507 · 3 | 4,068 | |
| 24 | Muse Spark 1.3 (Max) | Meta | Proprietary | $0.10 | 459 | 503 | 397 · 2 | 10,796 | |
| 25 | MiMo-V2.6-Flash | Xiaomi | MIT | $0.14 | 454 | 484 | 435 · 4 | 7,068 | |
| 26 | Grok 4.6 (High) | xAI | Proprietary | $2.00 | 452 | 463 | 436 · 2 | 11,966 | |
| 27 | Claude Fable 5 (High) | Anthropic | Proprietary | $10.00 | 448 | 470 | 434 · 7 | 16,951 | |
| 28 | Kimi K3 (Max) | Moonshot AI | Kimi K3 license | $2.85 | 441 | 504 | 403 · 6 | 21,281 | |
| 29 | GLM-5.3 (Max) |   Z.ai | MIT | $1.20 | 439 | 469 | 420 · 6 | 11,150 | |
| 30 | Muse Spark 1.3 (xHigh) | Meta | Proprietary | $0.10 | 438 | 466 | 397 · 2 | 10,243 | |
| 31 | Gemini 3.8 Flash (High) | Google | Proprietary | $0.75 | 433 | 431 | 435 · 3 | 12,342 | |
| 32 | Qwen3.8-Max (0902) |  Alibaba | Proprietary | $1.65 | 432 | 524 | 377 · 5 | 17,936 | |
| 33 | Qwen3.8-Max | Alibaba | Proprietary | $1.65 | 430 | 520 | 377 · 5 | 3,406 | |
| 34 | DeepSeek-V4-Flash (High) | DeepSeek | MIT | – | 430 | 430 | – | 4,927 | |
| 34 | DeepSeek-V4-Pro (High, 20260813) | DeepSeek | MIT | – | 430 | 430 | – | 4,676 | |
| 35 | Step 5 Preview (High) | StepFun | Proprietary | – | 420 | 420 | – | 5,966 | |
| 36 | Hy4 preview |  Tencent | Apache 2.0 | – | 412 | 477 | 373 · 7 | 4,765 | |
| 37 | GPT-5.6 Terra (xHigh, Codex Harness) | OpenAI | Proprietary | $2.00 | 411 | 375 | 436 · 6 | 15,388 | |
| 38 | Gemini 3.7 Flash (High) | Google | Proprietary | $0.75 | 405 | 437 | 385 · 3 | 14,473 | |
| 39 | GPT-6 Luna (Max) | OpenAI | Proprietary | $0.10 | 404 | 427 | 370 · 2 | 27,808 | |
| 40 | GLM-5.3-Flash | Z.ai | MIT | $0.15 | 403 | 453 | 371 · 3 | 21,587 | |
| 41 | GPT-5.6 Luna (xHigh, Codex Harness) | OpenAI | Proprietary | $0.20 | 401 | 374 | 420 · 6 | 12,522 | |
| 42 | GLM-5.2 (Max) | Z.ai | MIT | $0.75 | 397 | 452 | 363 · 9 | 17,817 | |
| 43 | Grok 4.5 | xAI | Proprietary | $2.00 | 397 | 404 | 392 · 6 | 11,018 | |
| 44 | Qwen3.8-Flash-Next | Alibaba | qwen-community-1.0 | – | 395 | 475 | 348 · 4 | 9,703 | |
| 45 | Claude Opus 4.8 (High) | Anthropic | Proprietary | $5.00 | 385 | 406 | 372 · 9 | 18,408 | |
| 46 | Claude Opus 4.8 | Anthropic | Proprietary | $5.00 | 378 | 389 | 372 · 9 | 17,323 | |
| 47 | Seed2.1 Pro Preview |  ByteDance | Proprietary | – | 374 | 374 | – | 16,294 | |
| 48 | Muse Spark 1.2 (xHigh) | Meta | Proprietary | $0.10 | 374 | 387 | 356 · 2 | 2,155 | |
| 49 | Muse Spark 1.1 | Meta | Proprietary | $1.25 | 374 | 394 | 361 · 4 | 7,935 | |
| 50 | Claude Sonnet 5 (High) | Anthropic | Proprietary | $2.00 | 370 | 392 | 356 · 7 | 16,832 | |
| 51 | GPT-5.5 (xHigh, Codex Harness) | OpenAI | Proprietary | $5.00 | 365 | 368 | 362 · 6 | 15,553 | |
| 52 | Muse Spark Open (Max) | Meta | muse-model-license | $1.25 | 364 | 380 | 353 · 4 | 1,721 | |
| 53 | Claude Opus 4.7 (High) | Anthropic | Proprietary | $5.00 | 363 | 408 | 336 · 6 | 17,859 | |
| 53 | Qwen3.8-27B | Alibaba | Apache 2.0 | $0.40 | 363 | 437 | 319 · 4 | 18,752 | |
| 54 | Claude Opus 4.7 | Anthropic | Proprietary | $5.00 | 363 | 406 | 336 · 6 | 18,167 | |
| 55 | GPT-5.5 (High, Codex Harness) | OpenAI | Proprietary | $5.00 | 356 | 348 | 362 · 6 | 17,853 | |
| 56 | Gemini 3.6 Flash (High) | Google | Proprietary | $1.50 | 356 | 390 | 334 · 3 | 8,722 | |
| 57 | Claude Opus 4.6 (High) | Anthropic | Proprietary | $5.00 | 347 | 398 | 315 · 5 | 19,316 | |
| 58 | GPT-5.5 (Codex Harness) | OpenAI | Proprietary | $5.00 | 346 | 324 | 362 · 6 | 15,811 | |
| 59 | Spark-X2.6 | Synlan | Proprietary | – | 344 | 344 | – | 1,210 | |
| 60 | Qwen3.6-Max-Preview | Alibaba | Proprietary | – | 344 | 344 | – | 2,700 | |
| 61 | Claude Opus 4.6 | Anthropic | Proprietary | $5.00 | 344 | 390 | 315 · 5 | 20,576 | |
| 62 | Gemini 3.5 Flash (High) | Google | Proprietary | $1.50 | 343 | 358 | 333 · 4 | 9,331 | |
| 63 | Gemini 3.5 Flash (Medium) | Google | Proprietary | $1.50 | 340 | 351 | 333 · 4 | 12,075 | |
| 64 | Kimi K2.6 | Moonshot AI | Modified MIT | $0.75 | 332 | 366 | 311 · 7 | 10,119 | |
| 65 | Qwen3.7-Max (20260517) | Alibaba | Proprietary | $1.25 | 331 | 371 | 306 · 8 | 8,855 | |
| 66 | DeepSeek-V4-Pro Preview (High) | DeepSeek | MIT | – | 330 | 330 | – | 13,087 | |
| 67 | Hy3 | Tencent | Apache 2.0 | $0.14 | 328 | 363 | 306 · 8 | 6,844 | |
| 68 | GLM-5.1 | Z.ai | MIT | $1.05 | 328 | 365 | 305 · 5 | 10,647 | |
| 68 | Kimi K2.7 Code | Moonshot AI | Modified MIT | $0.68 | 328 | 337 | 322 · 4 | 4,950 | |
| 69 | GPT-5.4 (High, Codex Harness) | OpenAI | Proprietary | $2.50 | 328 | 330 | 327 · 5 | 1,345 | |
| 70 | Claude Sonnet 4.6 | Anthropic | Proprietary | $3.00 | 325 | 377 | 292 · 4 | 22,733 | |
| 71 | MiniMax-M3 | MiniMax | MiniMax Community License | $0.28 | 323 | 345 | 308 · 7 | 21,032 | |
| 72 | GPT-5.4 (Medium, Codex Harness) | OpenAI | Proprietary | $2.50 | 322 | 314 | 327 · 5 | 1,374 | |
| 73 | Claude Opus 4.5 (High, 32K, 20251101) | Anthropic | Proprietary | – | 318 | 353 | 296 · 3 | 10,981 | |
| 74 | MiMo-V2.5-Pro | Xiaomi | MIT | $0.43 | 318 | 343 | 301 · 4 | 22,604 | |
| 75 | DeepSeek-V4-Pro | DeepSeek | MIT | – | 317 | 317 | – | 13,931 | |
| 76 | Claude Opus 4.5 (20251101) | Anthropic | Proprietary | – | 311 | 334 | 296 · 3 | 13,362 | |
| 77 | Gemini 3.1 Pro Preview | Google | Proprietary | $2.00 | 310 | 317 | 305 · 7 | 24,129 | |
| 78 | GPT-5.4 | OpenAI | Proprietary | $2.50 | 309 | 284 | 327 · 5 | 4,499 | |
| 79 | Qwen3.6-Plus | Alibaba | Proprietary | $0.50 | 308 | 328 | 295 · 8 | 18,511 | |
| 80 | DeepSeek-V4-Flash Preview (High) | DeepSeek | MIT | – | 305 | 305 | – | 10,074 | |
| 81 | MiMo-V2.5 | Xiaomi | MIT | $0.17 | 304 | 311 | 300 · 3 | 14,837 | |
| 82 | Gemini 3 Pro | Google | Proprietary | – | 303 | 312 | 290 · 2 | 14,054 | |
| 83 | MiMo-V2-Pro | Xiaomi | Proprietary | – | 301 | 307 | 297 · 4 | 7,225 | |
| 84 | Inkling-Small | Thinky | Apache 2.0 | $0.30 | 300 | 289 | 307 · 5 | 5,295 | |
| 85 | GLM-5 | Z.ai | MIT | $1.00 | 299 | 309 | 294 · 3 | 7,465 | |
| 86 | Beam 1.0 501B A23B | Reflection | Apache 2.0 | $0.40 | 299 | 309 | 292 · 8 | 2,932 | |
| 87 | Inkling | Thinky | Apache 2.0 | $0.95 | 298 | 294 | 301 · 3 | 17,073 | |
| 88 | Kimi K2.5 (Thinking) | Moonshot AI | Modified MIT | – | 297 | 309 | 289 · 6 | 20,008 | |
| 89 | Gemini 3 Flash | Google | Proprietary | $0.50 | 294 | 312 | 283 · 3 | 10,845 | |
| 90 | Gemini 3.5 Flash-Lite | Google | Proprietary | $0.30 | 294 | 307 | 275 · 2 | 236 | |
| 91 | GPT-5.2 | OpenAI | Proprietary | $1.75 | 293 | 295 | 291 · 2 | 1,082 | |
| 92 | MiniMax-M2.7 | MiniMax | Modified MIT | $0.30 | 292 | 282 | 299 · 4 | 15,890 | |
| 93 | GPT-5.3-Codex (Codex Harness) | OpenAI | Proprietary | $1.75 | 289 | 263 | 329 · 2 | 2,639 | |
| 94 | Kimi K2.5 (Instant) | Moonshot AI | Modified MIT | – | 288 | 287 | 289 · 6 | 3,246 | |
| 95 | Qwen3.5-397B-A17B | Alibaba | Apache 2.0 | $0.45 | 287 | 283 | 289 · 3 | 21,096 | |
| 96 | GPT-5.4 mini (High) | OpenAI | Proprietary | $0.75 | 285 | 281 | 287 · 3 | 11,893 | |
| 97 | GLM-4.7 | Z.ai | MIT | $0.40 | 284 | 308 | 269 · 4 | 3,948 | |
| 98 | GLM-5V-Turbo | Z.ai | Proprietary | – | 284 | 284 | – | 2,394 | |
| 99 | MiniMax-M2.5 | MiniMax | Modified MIT | $0.30 | 283 | 273 | 299 · 2 | 7,594 | |
| 100 | NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4) |  NVIDIA | OpenMDW-1.1 | – | 282 | 282 | – | 3,193 | |
| 101 | GPT-5 (Medium) | OpenAI | Proprietary | – | 282 | 295 | 273 · 3 | 3,089 | |
| 102 | GPT-5.1 (Medium) | OpenAI | Proprietary | $1.25 | 280 | 280 | – | 4,891 | |
| 103 | Claude Sonnet 4.5 (High, 32K, 20250929) | Anthropic | Proprietary | $3.00 | 279 | 279 | – | 13,213 | |
| 103 | Gemini 3 Flash (Minimal) | Google | Proprietary | $0.50 | 279 | 272 | 283 · 3 | 22,080 | |
| 104 | Claude Opus 4.1 (20250805) | Anthropic | Proprietary | – | 276 | 276 | 276 · 2 | 6,904 | |
| 105 | Solar Pro 4 | Upstage | Proprietary | $0.30 | 274 | 263 | 281 · 3 | 4,630 | |
| 106 | Claude Sonnet 4.5 (20250929) | Anthropic | Proprietary | $3.00 | 273 | 273 | – | 15,679 | |
| 107 | MiniMax-M2.1 Preview | MiniMax | MIT | – | 272 | 272 | – | 7,573 | |
| 108 | Grok 4.20 Beta (Reasoning, 0309) | xAI | Proprietary | – | 266 | 266 | – | 15,404 | |
| 109 | Qwen3.5-122B-A10B | Alibaba | Apache 2.0 | $0.29 | 266 | 256 | 281 · 2 | 8,210 | |
| 110 | GPT-5.2-Codex | OpenAI | Proprietary | $1.75 | 265 | 242 | 303 · 2 | 6,705 | |
| 111 | Qwen3.5-27B | Alibaba | Apache 2.0 | $0.26 | 263 | 255 | 275 · 2 | 7,767 | |
| 112 | Gemma 4 31B | Google | Apache 2.0 | $0.09 | 262 | 262 | – | 15,489 | |
| 113 | GPT-5.1-Codex | OpenAI | Proprietary | – | 258 | 242 | 286 · 2 | 4,981 | |
| 114 | DeepSeek-V3.2 (Thinking) | DeepSeek | MIT | $0.28 | 258 | 258 | – | 6,860 | |
| 115 | Gemma 4 26B A4B | Google | Apache 2.0 | $0.07 | 255 | 255 | – | 1,494 | |
| 116 | MiMo-V2-Flash (Non-Thinking) | Xiaomi | MIT | – | 255 | 237 | 267 · 4 | 5,645 | |
| 117 | GLM-4.6 | Z.ai | MIT | $0.50 | 255 | 244 | 272 · 2 | 6,755 | |
| 118 | Grok 4.3 | xAI | Proprietary | $1.25 | 255 | 255 | – | 14,906 | |
| 119 | Hy3 preview | Tencent | tencent-hunyuan-community | – | 254 | 254 | – | 1,444 | |
| 120 | Muse Glimmer | Meta | Apache-2.0 | – | 253 | 253 | – | 1,632 | |
| 121 | Claude Haiku 4.5 (20251001) | Anthropic | Proprietary | $1.00 | 252 | 237 | 274 · 2 | 28,961 | |
| 122 | MiniMax-M2 | MiniMax | Apache 2.0 | $0.30 | 251 | 217 | 275 · 5 | 6,683 | |
| 123 | Laguna M.1 |  Poolside | Apache 2.0 | – | 249 | 249 | – | 5,359 | |
| 124 | MiMo-V2-Flash (Thinking) | Xiaomi | MIT | – | 245 | 214 | 267 · 4 | 1,681 | |
| 125 | GPT-5.1 | OpenAI | Proprietary | $1.25 | 244 | 244 | – | 10,675 | |
| 126 | DeepSeek-V3.2-Exp | DeepSeek | MIT | – | 240 | 203 | 267 · 5 | 3,968 | |
| 127 | Qwen3-Coder-480B-A35B-Instruct | Alibaba | Apache 2.0 | $0.30 | 235 | 205 | 257 · 4 | 12,675 | |
| 128 | DeepSeek-V3.2 | DeepSeek | MIT | $0.28 | 234 | 234 | – | 9,327 | |
| 129 | Kimi K2 Thinking Turbo | Moonshot AI | Modified MIT | – | 232 | 232 | – | 13,045 | |
| 130 | Qwen3.5-35B-A3B | Alibaba | Apache 2.0 | $0.14 | 222 | 194 | 271 · 2 | 1,612 | |
| 131 | Laguna XS.2 | Poolside | Apache 2.0 | – | 221 | 221 | – | 4,204 | |
| 132 | Gemini 2.5 Pro | Google | Proprietary | $1.25 | 211 | 179 | 265 · 2 | 2,724 | |
| 133 | Mistral Medium 3.5 | Mistral | Modified MIT | $1.50 | 199 | 199 | – | 2,299 | |
| 134 | Gemini 3.1 Flash-Lite Preview | Google | Proprietary | $0.25 | 195 | 195 | – | 14,350 | |
| 135 | KAT-Coder-Pro V1 | Unknown | Proprietary | – | 195 | 195 | – | 1,503 | |
| 136 | Mercury 2 |   Inception | Proprietary | $0.25 | 192 | 153 | 262 · 2 | 954 | |
| 137 | GPT-5.1-Codex-Mini | OpenAI | Proprietary | – | 188 | 188 | – | 1,154 | |
| 137 | Qwen3.5-Flash | Alibaba | Proprietary | – | 188 | 188 | – | 1,321 | |
| 138 | Grok 4.1 Fast (Reasoning) | xAI | Proprietary | $0.20 | 187 | 187 | – | 5,473 | |
| 139 | Trinity Large Thinking | Unknown | Apache 2.0 | – | 185 | 185 | – | 1,352 | |
| 140 | Mistral Large 3 | Mistral | Apache 2.0 | $0.50 | 182 | 182 | – | 838 | |
| 141 | Grok 4.1 (Thinking) | xAI | Proprietary | – | 174 | 174 | – | 961 | |
| 142 | Devstral 2 | Mistral | Modified MIT | – | 165 | 165 | – | 1,182 | |
| 143 | Granite 4.1 8B |  IBM | Apache 2.0 | – | 163 | 163 | – | 1,766 | |
| 144 | Grok Code Fast 1 | xAI | Proprietary | $0.20 | 151 | 151 | – | 789 | |
| 145 | Grok 4 Fast (Reasoning) | xAI | Proprietary | $0.20 | 148 | 148 | – | 742 | |
| 146 | Devstral Medium (2507) | Mistral | Proprietary | – | 115 | 115 | – | 842 |
Leaders by app type
The top model for each kind of task, scored from votes on that task alone, and how many models were ranked.
| Task | Leader | Score | Models ranked |
|---|---|---|---|
| HTML apps | GPT-6.1 Sol (Max) OpenAI | 659 | 150 |
| React apps | Claude Opus 5.5 (Max) Anthropic | 727 | 135 |
| Web apps | Claude Opus 5.5 (Max) Anthropic | 675 | 151 |
Related benchmarks
Fixed tests with right and wrong answers, scored out of 100%.
SWE-Bench Verified
Fixing real bugs in real GitHub projects, checked by the projects' own tests. · Top 10 of 123 · Higher is better
SWE-Bench Pro
Harder, longer software engineering tasks from real codebases. · Top 10 of 63 · Higher is better
Terminal-Bench 2.0
Getting real work done in a command-line terminal. · Top 10 of 54 · Higher is better