1 Best pickResearched 1d ago
Q/$ B
SWE-bench Pro
52.6%
#34 of 99 priced candidates
Source: SWE-bench Pro score 52.6 plus DeepSeek Platform pricing drives this Q/$ rank.
Updated today · 1,687 models · 130 providers
Shortlist for coding / agent workflows
Sorted for $5/$15 out; needs 200K context, tools.
SWE-bench Pro
52.6%
Source: SWE-bench Pro score 52.6 plus DeepSeek Platform pricing drives this Q/$ rank.
HumanEval
68.4%
Source: HumanEval score 68.4 plus DeepInfra pricing drives this Q/$ rank.
HumanEval
88.4%
Source: HumanEval score 88.4 plus SiliconFlow pricing drives this Q/$ rank.
This week: 161 new models · 62 price cuts · 13 benchmark refreshes