DeepSWE (version not stated)
Agentic software engineering
codingactive
Score in %
Definition
- Version
- not in the notes
- Category
- coding
- Maintainer
- not in the notes
- Status
- active
- Home page
- not in the notes
Sources
- Gemini 3.6 Flash in Antigravity: Agentic Coding Guide noqta.tn https://noqta.tn/en/blog/gemini-3-6-flash-antigravity-agentic-coding-2026
- Kimi K3 Open Weights Ship: 2.8T Parameters, 1.4 TB to Run rits.shanghai.nyu.edu https://rits.shanghai.nyu.edu/ai/kimi-k3-open-weights-ship-2-8t-parameters-1-4-tb-to-run/
- huggingface.co/moonshotai/Kimi-K3 https://huggingface.co/moonshotai/Kimi-K3
- huggingface.co/XiaomiMiMo https://huggingface.co/XiaomiMiMo
- huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL
Scores at a Glance
Score in %
Higher is better for this unit.
1
GPT-5.6 SolOpenAI
MiMo-V2.6-ProXiaomi
Kimi K3Moonshot AIKimi Code harness; 67.3 with mini-SWE-agent
Gemini 3.6 FlashGoogle DeepMindGemini 3.5 FlashGoogle DeepMind
GPT-5.6 SolOpenAI73 %
vendor-reported2MiMo-V2.6-ProXiaomi71.9 %
vendor-reported3Kimi K3Moonshot AIKimi Code harness; 67.3 with mini-SWE-agent67.5 %
vendor-reported4Gemini 3.6 FlashGoogle DeepMind49 %
vendor-reported5Gemini 3.5 FlashGoogle DeepMind37 %
vendor-reported0255075100
Where this comes from
Each bar is one model, with its most independent recorded score. Hatched bars are vendor-reported; the word beside each bar says who ran it. Settings and dates differ between rows; the Scores table below has every result.
Scores
| Model | Score | Who ran it | Setting | View or population | Date | Source |
|---|---|---|---|---|---|---|
| GPT-5.6 Sol | 73 % | vendor-reported | – | – | – | rits.shanghai.nyu.edu |
| MiMo-V2.6-Pro | 71.9 % | vendor-reported | – | – | – | huggingface.co |
| Kimi K3 | 67.5 % | vendor-reported | Kimi Code harness; 67.3 with mini-SWE-agent | – | – | huggingface.co |
| Gemini 3.6 Flash | 49 % | vendor-reported | – | – | – | noqta.tn |
| Gemini 3.5 Flash | 37 % | vendor-reported | – | – | – | noqta.tn |
Where this comes from
Scores from different settings, harnesses or dates are not directly comparable; the row says which. Sorted by score, highest first, as a reading aid, not a ranking.
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "DeepSWE (Version Not Stated)", https://aurion.semai.id/benchmarks/deepswe/, records of October 2026.