Skip to content

Google · Vision LLM

Gemini 3.5 Flash on ResumeBench: 85.5 F1, rank 5 of 18

Gemini 3.5 Flash ranks 5th of 18 on ResumeBench: 85.5 F1, $35 per 1,000 pages, 43 seconds per resume. It is 1.7 points behind LlamaParse Extract Agentic Plus. LlamaParse Extract Agentic scores 85.5 for $23.

2nd of 11 vision LLMs. Strongest on references (+5.5 points), resume-builder templates (+5.2 points) and overlapping jobs (+4.7 points). Weakest on content as images (−11.5 points), wording traps (−4.1 points) and distractor layouts (−4.0 points).

Rank
5/18
F1
85.5
Precision
86.9
Recall
87.9
Perfect resumes
34/300
Per 1,000 pages
$35
Median per resume
43 s

Averaged over 300 resumes. Dates (scored separately): 97.8. 90% finish within 70 s.

Where Gemini 3.5 Flash wins and loses

F1 by kind of resume, against the median system.

Reading the page

Gemini 3.5 Flash by reading the page
Kind of resume# of resumesF1MedianGapRank
Handwritten950.351.5−1.213th
Image only2875.171.4+3.77th
Sidebar or grid8285.181.9+3.14th
Photo4585.184.0+1.07th

Distractors

Gemini 3.5 Flash by distractors
Kind of resume# of resumesF1MedianGapRank
Rotated & warped pages2187.482.9+4.57th
Hidden text & prompt injections1486.784.5+2.19th
Content as images876.788.2−11.515th
Wording traps1680.484.5−4.114th
Typos kept on purpose3784.483.7+0.79th

Extraction challenges

Gemini 3.5 Flash by extraction challenges
Kind of resume# of resumesF1MedianGapRank
Many entries8384.381.9+2.45th
Long bullet lists10183.981.4+2.46th
Split across pages8584.682.2+2.45th
Big skills list6086.083.6+2.41st
Custom sections10583.680.8+2.72nd
Combined headings3784.282.7+1.53rd
Promotions8887.383.5+3.92nd
Partial dates13884.983.4+1.55th
Overlapping jobs6986.481.7+4.71st
Accented text1386.485.5+0.88th

Noise

Gemini 3.5 Flash by noise
Kind of resume# of resumesF1MedianGapRank
Repeated header6786.385.2+1.26th
References5590.785.2+5.51st
Busy page1880.880.1+0.78th

Source

Gemini 3.5 Flash by source
Kind of resume# of resumesF1MedianGapRank
Everyday resumes21287.285.1+2.12nd
Expert-built (Canva / LaTeX)3284.683.6+1.05th
Resume-builder templates1086.481.2+5.22nd
Scanned & rotated1987.282.9+4.38th
Distractor layouts1881.485.3−4.014th

Layout

Gemini 3.5 Flash by layout
Kind of resume# of resumesF1MedianGapRank
Single column19585.184.0+1.16th
Two-column2390.086.8+3.22nd
Two-column sidebar5483.480.0+3.44th
Three-column2085.684.8+0.85th
Table grid895.094.1+0.97th

Length

Gemini 3.5 Flash by length
Kind of resume# of resumesF1MedianGapRank
Short14888.387.1+1.24th
Medium9085.084.7+0.48th
Long6279.274.6+4.52nd

By section of the resume

F1 by resume section, against the median system.

Gemini 3.5 Flash by resume section
SectionResumesF1MedianGap
Contact details30095.393.5+1.7
Experience30086.586.3+0.2
Education29791.589.4+2.1
Summary29089.381.7+7.6
Certifications24584.081.1+2.8
Awards9280.375.4+4.9
Skills28993.592.1+1.4
Languages17084.384.3−0.0
Custom sections21650.149.9+0.2
Interests6394.291.7+2.4
Volunteering7983.784.5−0.8
Profiles3564.866.7−2.0
Projects8080.483.2−2.8
Publications5283.478.6+4.8

On the published resumes

Best and worst of the 20 public resumes.

Gemini 3.5 Flash on published resumes
ResumeWhat makes it hardF1
Lead Respiratory Therapist (Adult ICU)Long resume, 34 entries96.7
Senior Data Scientist (Lead)Fifty skills, seven pages95.8
Senior Manufacturing EngineerThird-person prose, no bullets93.4
Senior Game DesignerGame designer's Canva resume54.8
senior Consultant - package ImplementationHandwritten resume42.1
Senior ocularist and practice ownerHandwritten resume, six pages39.8

How it was run

Gemini API, Gemini 3.5 Flash, page images in, JSON schema out. Same PDF, schema and instructions as every other system, one run per resume (4 November 2026).

Gemini 3.5 Flash on Google’s site

Run the open-source harness: list its pipelines, or grade your own predictions (GitHub):

$ pip install "resume-bench[runners] @ git+https://github.com/careerflow/ResumeBench"
$ hf auth login   # the dataset is gated: accept its terms on Hugging Face first
$ resume-bench download
$ resume-bench providers
$ resume-bench grade-file my_predictions.jsonl --split test

Closest alternatives

Nearest in F1.

Embed this result

Shows its current rank. Updates with each run.

Gemini 3.5 Flash: #5 on ResumeBench resume extraction benchmark
<a href="https://www.resumebench.ai/systems/gemini-3-5-flash"><img src="https://www.resumebench.ai/badge/gemini-3-5-flash.svg" alt="Gemini 3.5 Flash: #5 on ResumeBench resume extraction benchmark" height="28"></a>

Choosing a resume extraction model

All 18 models · Leaderboard

Frequently asked questions (FAQs)

How accurate is Gemini 3.5 Flash at resume extraction?

85.5 F1 on ResumeBench's 300 resumes, 5th of 18 models.

How much does Gemini 3.5 Flash cost to parse resumes?

$35 per 1,000 pages, about $0.098 per resume.

How fast is Gemini 3.5 Flash?

43 seconds per resume (median).

Which resumes is Gemini 3.5 Flash best at?

References, resume-builder templates and overlapping jobs. Weakest on content as images, wording traps and distractor layouts.

What is a cheaper alternative to Gemini 3.5 Flash?

LlamaParse Extract Agentic: 85.5 F1 for $23 per 1,000 pages.