DeepSWE leaderboard: which LLMs score highest
DeepSWE (long-horizon software engineering) currently has scores for 19 models. The top score is 73.8% by Gemini 3.8 Flash (high); the median model scores 67.0% and the lowest scores 30.5%.
| # | Model | Organisation | DeepSWE score | Cheapest input $/M | Source | Captured |
|---|---|---|---|---|---|---|
| 1 | Gemini 3.8 Flash (high) | Google DeepMind | 73.8% | $0.750 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 2 | Claude Opus 5 (max) | Anthropic | 73.6% | $5.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 3 | GPT-6 Astra (max) | OpenAI | 73.2% | $10.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 4 | GPT-5.6 Sol (max) | OpenAI | 72.7% | $2.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 5 | Claude Fable 5 (max) | Anthropic | 69.7% | $10.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 6 | GPT-5.6 Terra (max) | OpenAI | 69.6% | $2.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 7 | GLM-5.3 (max) | Z.ai (Zhipu AI) | 69.0% | $0.070 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 8 | Kimi K3 (max) | Moonshot | 68.5% | $1.29 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 9 | GPT-5.6 Luna (max) | OpenAI | 67.2% | $0.200 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 10 | GPT-5.5 (xhigh) | OpenAI | 67.0% | $5.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 11 | Grok 4.6 (xhigh) | xAI | 66.7% | $1.25 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 12 | Gemini 3.7 Flash (high) | Google DeepMind | 65.3% | $0.750 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 13 | GLM-5.3-Flash (max) | Z.ai (Zhipu AI) | 63.4% | $0.110 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 14 | Qwen3.8 Max (xhigh) | Alibaba | 57.5% | $1.65 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 15 | Muse Spark 1.2 (xhigh) | Meta AI | 54.9% | $1.25 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 16 | Claude Sonnet 5 (max) | Anthropic | 53.8% | $2.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 17 | Grok 4.5 (high) | xAI | 53.8% | $2.00 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 18 | GPT-5.4 (xhigh) | OpenAI | 51.8% | $2.50 | Epoch AI Benchmarking Hub | 2026-10-05 |
| 19 | Kimi K2.7 Code | Moonshot | 30.5% | $0.671 | Epoch AI Benchmarking Hub | 2026-10-05 |
Compare all benchmarks side by side on the live leaderboard →
Among the ten highest scorers, the cheapest listed API price belongs to GLM-5.3 (max) at $0.070 per million input tokens (score 69.0%).
What DeepSWE measures
DeepSWE measures long-horizon software engineering.
How to read these scores
Scores are percentages (higher is better): the share of questions or tasks the model completed correctly. Where a model is listed at several reasoning efforts, only its highest-effort row is shown. See the methodology for how rows are chosen.
Where the data comes from
Epoch AI Benchmarking Hub (CC BY 4.0). Scores were last captured 2026-10-05; the Captured column gives each row's date.