Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Loading page…
Preferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Coding benchmark
SWE-bench Verified (mini-swe-agent-v2). Every model the catalog carries a published SWE-bench Verified* value for, ranked by that value.
A display-only SWE-bench Verified reference from Arcee AI's Trinity-Large-Thinking comparison chart. 5 models in the ModelScale catalog have a published value.
Each point is a model with a published SWE-bench Verified* value and complete token prices, on the 75% input / 25% output blend. Frontier models are spaced evenly along the cost axis.
5 models with a published SWE-bench Verified* value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Agent scaffold benchmark |
|---|---|---|---|
| 1 | Anthropic | 75.6 | |
| 2 | MiniMax | 75.4 | |
| 3 | Z.AI | 72.8 |
What the benchmark measures and how its value is published, from the source’s own record.
SWE-bench Verified* (SWE-bench Verified (mini-swe-agent-v2)) is a coding benchmark published by Trinity-Large-Thinking: Scaling an Open Source Frontier Agent. A display-only SWE-bench Verified reference from Arcee AI's Trinity-Large-Thinking comparison chart. Its evaluation set is described by the source as repository task completion. The source rates its difficulty as professional software engineering.
Each model's result is published by Trinity-Large-Thinking: Scaling an Open Source Frontier Agent as agent scaffold benchmark. ModelScale prints that value to 1 decimal place, exactly as published, and never converts or rescales it. This page ranks the 5 models that carry a published value, highest value first; a model the source never scored is left out rather than ranked at zero. The source does not state whether a higher value is the better result, so neither does this page.
| 4 | Moonshot AI | 70.8 |
| 5 | Arcee AI | 63.2 |
Evidence key: Observed
Rows are ordered by the value Trinity-Large-Thinking: Scaling an Open Source Frontier Agent published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.