BlogLLM Benchmark Scores as Published, Side by Side

Gemini 3 Flash: Benchmark Scores as Published

Every result the catalog holds for Gemini 3 Flash: 15 tests in 8 areas, from 3 pages. 14 printed by Google itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.

Developer GoogleReleased 2025-12-17Context window 1.05MMax output 65.5KInput Text, image, video, audio, PDFOutput TextWeights ClosedData version 2026103001, as of 2026-10-30

Knowledge Work

TestValuePrinted bySetup as printed
GDPval-AA (Elo)1204Google, Gemini 3.5 Flash model cardchecked 2026-10-15-
Vals Finance Agent v242.6%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-

Agentic Coding

TestValuePrinted bySetup as printed
SWE-bench Verified78%Anthropic, Claude Sonnet 4.6checked 2026-10-15-
SWE-Bench Pro49.6%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-
Terminal-Bench 2.158%Google, Gemini 3.5 Flash model cardchecked 2026-10-15Harness: Terminus-2

Long Context

TestValuePrinted bySetup as printed
MRCR v2 8-needle 128K67.2%Google, Gemini 3.5 Flash model cardchecked 2026-10-15Context: 128k (average)
MRCR v2 8-needle 1M (pointwise)22.1%Google, Gemini 3.5 Flash model cardchecked 2026-10-15Context: 1M (pointwise)

Computer Use

TestValuePrinted bySetup as printed
OSWorld-Verified65.1%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-

Multimodal Understanding

TestValuePrinted bySetup as printed
CharXiv Reasoning (no tools)80.3%Google, Gemini 3.5 Flash model cardchecked 2026-10-15Tools: none
MMMU-Pro81.2%Google, Gemini 3.5 Flash model cardchecked 2026-10-15Tools: none

General Knowledge and Reasoning

TestValuePrinted bySetup as printed
Humanity's Last Exam (no tools)33.7%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-
ARC-AGI-233.6%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-

Tool Use and Workflows

TestValuePrinted bySetup as printed
Toolathlon49.4%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-
MCP Atlas62%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-

Engineering and Design

TestValuePrinted bySetup as printed
Blueprint-Bench 2 (normalized score)0%Google, Gemini 3.5 Flash model cardchecked 2026-10-15-

Price

API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked.

PriceValuePrinted bySetup as printed
Input$0.5Google, Gemini API pricingchecked 2026-10-15-
Cached input$0.05Google, Gemini API pricingchecked 2026-10-15-
Output$3Google, Gemini API pricingchecked 2026-10-15-
Input, batch$0.25Google, Gemini API pricingchecked 2026-10-15-
Cached input, batch$0.05Google, Gemini API pricingchecked 2026-10-15-
Output, batch$1.5Google, Gemini API pricingchecked 2026-10-15-
Input, flex$0.25Google, Gemini API pricingchecked 2026-10-15-
Cached input, flex$0.05Google, Gemini API pricingchecked 2026-10-15-
Output, flex$1.5Google, Gemini API pricingchecked 2026-10-15-
Input, priority$0.9Google, Gemini API pricingchecked 2026-10-15-
Cached input, priority$0.09Google, Gemini API pricingchecked 2026-10-15-
Output, priority$5.4Google, Gemini API pricingchecked 2026-10-15-

More from Google

Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.