Benchmarks · coding
SciCode
SciCode: implementing scientific research code from paper descriptions; sub-problem accuracy. Artificial Analysis marks this evaluation under review, so its scores may change. The reason is SciCode-Verified (arXiv 2608.04975), an unreviewed preprint from the authors of the proposed fixes, which reports dataset defects that reject correct solutions. AA removed SciCode from its Intelligence Index in v5.0.
99 models · 182 results · data as of 2026-11-06 · artificialanalysis.ai ↗
Leaderboard
Opus 5.5 max
More from this benchmark
3 published measuresScore vs any published measure3 measures: sub-scores, splits, counts
76 more models under All models in settings.
Breakdown
darker is better within each column · click a column to sort| Model | |||
|---|---|---|---|
| Opus 5.5 max | 1,222,503 | 366,615 | 6,144,155 |
| Fable 5.1 max | 1,196,326 | 257,078 | 3,296,726 |
| Gemini 4 Argon high | 1,437,250 | 112,988 | 2,643,094 |
| Fable 5 max, fallback | 1,109,989 | 202,153 | 1,181,023 |
| Sonnet 5.5 max | 1,170,592 | 273,839 | 7,498,720 |
| MiMo V2.6 Pro | 1,095,139 | 380,305 | 1,652,819 |
| Kimi K3.1 max | 656,500 | 106,176 | 1,773,698 |
| Muse Wonder (Max) | 767,980 | 161,222 | 1,629,083 |
| Gemini 3.7 Flash medium | 672,246 | 112,728 | 522,976 |
| Muse Spark 1.3 xhigh | 795,883 | 168,116 | 2,116,326 |
| Kimi K3 max | 706,608 | 134,340 | 879,539 |
| GLM 5.3 max | 1,020,664 | 201,454 | 4,842,757 |
| Muse Spark 1.1 xhigh | 836,006 | 187,379 | 1,646,216 |
| Gemini 3.1 Pro | 722,635 | 135,983 | 1,734,728 |
| Muse Spark Open xhigh | 853,129 | 207,084 | 2,284,507 |
| GPT-5.6 Sol high | 690,024 | 123,084 | 247,031 |
| Grok 4.7 high | 659,420 | 107,513 | 4,439,722 |
| GPT-6 Sol max | 565,543 | 76,470 | 1,034,130 |
| Muse Spark 1.2 xhigh | 969,555 | 241,431 | 3,571,783 |
| Gemini 3.8 Flash high | 720,264 | 129,595 | 5,625,621 |
| GPT-6 Astra max | 660,069 | 115,195 | 1,090,604 |
| Grok 4.6 high | 653,668 | 101,310 | 985,361 |
| Opus 5 max | 1,158,185 | 249,702 | 874,716 |
| GPT-5.5 high | 745,068 | 143,681 | 759,843 |
| DeepSeek V4.1 Flash 1017 max | 709,919 | 137,217 | 3,559,853 |
All results
182 rows| # | Model | Effort | SciCode |
|---|---|---|---|
| 1 | Opus 5.5 | max | 66.9% |
| 2 | Opus 5.5 | xhigh | 65.0% |
| 3 | Fable 5.1 | max | 63.1% |
| 4 | Gemini 4 Argon | high | 61.8% |
| 5 | Fable 5 | max | 61.0% |
| 6 | Sonnet 5.5 | max | 61.0% |
| 7 | Fable 5.1 | xhigh | 60.9% |
| 8 | MiMo V2.6 Pro | 60.9% | |
| 9 | Kimi K3.1 | max | 60.8% |
| 10 | Opus 5.5 | high | 60.4% |
| 11 | Muse Wonder (Max) | max | 60.3% |
| 12 | Gemini 3.7 Flash | medium | 59.8% |
| 13 | Muse Spark 1.3 | xhigh | 59.7% |
| 14 | Kimi K3 | max | 59.5% |
| 15 | Opus 5.5 | medium | 59.3% |
| 16 | GLM 5.3 | max | 59.0% |
| 17 | Muse Spark 1.1 | xhigh | 58.8% |
| 18 | Muse Spark 1.3 | max | 58.8% |
| 19 | Fable 5.1 | high | 58.7% |
| 20 | Gemini 3.1 Pro | 58.7% | |
| 21 | Opus 5.5 | low | 58.6% |
| 22 | Muse Spark Open | xhigh | 58.4% |
| 23 | GPT-5.6 Sol | high | 57.8% |
| 24 | Grok 4.7 | high | 57.8% |
| 25 | GPT-6 Sol | max | 57.6% |
| 26 | GPT-5.6 Sol | medium | 57.4% |
| 27 | Grok 4.7 | xhigh | 57.4% |
| 28 | Muse Spark 1.2 | xhigh | 57.4% |
| 29 | Sonnet 5.5 | xhigh | 57.3% |
| 30 | Gemini 3.7 Flash | high | 57.2% |