SWE-bench Verified benchmark
31 AI models ranked on SWE-bench Verified. Each row shows the model’s score next to its cheapest API price per 1M tokens where it has one, so you can weigh quality against cost.
Top model today
DeepSeek V4 Pro (DeepSeek) leads at 81%, from $0.43 per 1M input.
Data via SWE-bench →Models ranked
31
on this benchmark
Top score
81%
current leader
Type
Single eval
benchmark
Updated
15 July 2026
last source fetch
Cost vs SWE-bench Verified
19 priced modelsEach labelled name links to that model’s page. Curated to the cost-efficiency frontier, top scorers and cheapest so labels stay readable; the full field is in the leaderboard table below (and via “select all”). Reasoning models (extended “thinking” before answering) carry a ring around the dot.
Models ranked on SWE-bench Verified
highest score first · 31 models| # | ||||||
|---|---|---|---|---|---|---|
| 1 | DeepSeek V4 Pro |  DeepSeek | 81% | $0.43 | $0.87 | |
| 2 | MiniMax M3 |  MiniMax | 81% | $0.30 | $1.20 | |
| 3 | Kimi K2.6 |  Moonshot | 80% | n/a | n/a | |
| 4 | DeepSeek V4 Flash | DeepSeek | 79% | n/a | n/a | |
| 5 | MiMo-V2.5-Pro |  Xiaomi | 79% | $1.00 | $3.00 | |
| 6 | Hy3-preview | Tencent | 78% | n/a | n/a | |
| 7 | GLM 5 |  Zhipu | 78% | $0.80 | $2.56 | |
| 8 | Mistral Medium 3.5 128B |  Mistral | 78% | $1.50 | $7.50 | |
| 9 | Qwen3.6 27B |  Alibaba | 77% | $0.15 | $0.50 | |
| 10 | Qwen3.5 397B A17B | Alibaba | 76% | $0.60 | $3.60 | |
| 11 | MiniMax M2.5 | MiniMax | 76% | n/a | n/a | |
| 12 | Hy3-preview | Tencent | 74% | n/a | n/a | |
| 13 | Step 3.5 Flash 2603 | Stepfun | 74% | n/a | n/a | |
| 14 | MiniMax M2.1 | MiniMax | 74% | $0.27 | $1.20 | |
| 15 | Ring-2.6-1T | Inclusionai | 74% | n/a | n/a | |
| 16 | GLM 4.7 | Zhipu | 74% | $0.40 | $1.50 | |
| 17 | Qwen3.6 35B A3B | Alibaba | 73% | $0.14 | $0.45 | |
| 18 | Qwen3.5-27B | Alibaba | 72% | $0.30 | $2.40 | |
| 19 | Ling-2.6-1T | Inclusionai | 72% | n/a | n/a | |
| 20 | Qwen3.5-122B-A10B | Alibaba | 72% | $0.25 | $1.75 | |
| 21 | Nemotron 3 Ultra |  NVIDIA | 72% | n/a | n/a | |
| 22 | Kimi K2 Thinking | Moonshot | 71% | $0.60 | $1.20 | |
| 23 | Kimi K2.5 | Moonshot | 71% | $0.50 | $2.80 | |
| 24 | Qwen3 Coder Next | Alibaba | 71% | $0.50 | $1.20 | |
| 25 | MiniMax M2 | MiniMax | 69% | $0.26 | $1.02 | |
| 26 | North Mini Code |  Cohere | 68% | n/a | n/a | |
| 27 | gpt-oss-120b |  OpenAI | 62% | $0.05 | $0.25 | |
| 28 | Ling 2.6 Flash | Inclusionai | 61% | n/a | n/a | |
| 29 | gpt-oss-20b | OpenAI | 61% | $0.01 | $0.07 | |
| 30 | Nemotron 3 120B A12b | NVIDIA | 60% | $0.50 | $1.50 | |
| 31 | GLM 4.7 Flash | Zhipu | 59% | $0.00 | $0.00 |
Frequently asked questions
What is the SWE-bench Verified benchmark?
SWE-bench Verified is a 500-task subset of SWE-bench, human-validated by OpenAI to remove unsolvable or under-specified problems. Each task is a real GitHub issue from a popular Python repository, and the model must produce a patch that makes the project's hidden test suite pass. Scored as the percentage of issues resolved.
Which AI model scores highest on SWE-bench Verified?
DeepSeek V4 Pro (DeepSeek) leads with 81%, from $0.43 per 1M input tokens.
How many models are ranked on SWE-bench Verified?
31 models carry a SWE-bench Verified score. Scores come from Hugging Face leaderboards; where a model has a live API price it is verified against LiteLLM and OpenRouter, and models with no current price are listed without one.
Other benchmarks
compare the same models on a different evalEvery weekday
AI moves fast. Here's your debrief.
News, analysis, tools, and more.
For people who build with AI