Skip to content

Which resume extraction model handles resumes with many entries best?

83 of the 300 resumes are resumes with many entries. The typical system scores 2.9 points lower on them. Best: Extend Max Context, 85.6 F1. Gemini 3.5 Flash is within two points for $35 per 1,000 pages.

Resumes with 18 or more jobs, schools, projects and other entries, so the output is long and every entry has to be found. Best document API: Extend Max Context (85.6). Best vision LLM: GPT-5.6 (84.5). Extend Max Context ranks 6 places higher here than overall.

Every system on resumes with many entries

F1 on the 83 resumes with many entries, against each system’s F1 on all 300 resumes. Cost and time are for all resumes.

Ranking on resumes with many entries
ModelF1 hereF1 overallGapPer 1,000 pagesMedian time
Extend Max Context85.684.7+0.9$5043 s
LlamaParse Extract Agentic Plus85.087.2−2.1$5470 s
GPT-5.684.584.7−0.3$7438 s
Reducto Deep Extract84.385.6−1.2$4065 s
Gemini 3.5 Flash84.385.5−1.2$3543 s
GPT-5.583.883.5+0.3$8748 s
Claude Opus 4.883.785.5−1.8$9031 s
LlamaParse Extract Agentic83.085.5−2.5$2355 s
Qwen 3.8 Flash82.184.1−2.0$2.98122 s
Qwen 3.8 27B81.683.8−2.1$9.5936 s
Extend Extract80.281.0−0.7$1231 s
LlamaParse Extract Cost-Effective79.683.5−3.9$8.1829 s
Reducto Extract79.581.7−2.2$2016 s
Qwen 3.5 35B78.977.9+1.0$3.9178 s
Qwen 3.5 9B73.976.0−2.1$1.31138 s
GPT-5.4 Nano59.761.9−2.2$2.7620 s
Gemma 4 26B58.062.8−4.8$0.5629 s
GLM 5.3 Flash43.159.2−16.2$0.8046 s

What to do about it

Long careers and academic CVs. Weaker systems skip entries.

Published examples

12 public resumes of this kind.

Other hard resumes

Frequently asked questions (FAQs)

Which resume extraction model is best for resumes with many entries?

Extend Max Context, at 85.6 F1. Gemini 3.5 Flash is within two points for $35 per 1,000 pages.

How hard are resumes with many entries for parsers?

The typical model scores 2.9 points lower on them than on other resumes.

How many resumes with many entries are in ResumeBench?

83 of the 300.