Model card
Gemini 3 Flash.
GoogleapiUndisclosed params
§ 02 · Benchmarks
Every benchmark Gemini 3 Flash has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | LiveCodeBench | Computer Code · Code Generation | pass@1 | 90.8% | #2 | 2026-03-15 | source ↗ |
| 02 | ParseBench | Computer Vision · Document Parsing | accuracy | 71.0% | #3 | — | source ↗ |
| 03 | GPQA Diamond | Reasoning · Multi-step Reasoning | accuracy | 90.4% | #4 | — | source ↗ |
| 04 | SWE-Bench Verified | Computer Code · Code Generation | resolve-rate | 78.0% | #8 | 2025-12-17 | source ↗ |
| 05 | SWE-bench Verified | Agentic AI · SWE-bench | resolve-rate | 78.0% | #9 | — | source ↗ |
| 06 | SWE-bench | Computer Code · Code Generation | resolve-rate-agentic | 75.8% | #13 | 2026-02-01 | |
| 07 | SWE-bench | Computer Code · Code Generation | resolve-rate | 75.8% | #14 | 2026-02-01 | source ↗ |
| 08 | MMLU | Reasoning · Commonsense Reasoning | accuracy | 89.6% | #18 | 2026-01-01 | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area
Where Gemini 3 Flash actually performs.
§ 04 · Papers
1 paper with results for Gemini 3 Flash.
- 2023-10-10· Computer Code· 1 result
SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
Carlos E. Jimenez, John Yang, Alexander Wettig, Shunyu Yao et al.
§ 05 · Related models
Other Google models scored on Codesota.
§ 06 · Sources & freshness
Where these numbers come from.
google-blog
2
results
vendor
1
result
blog-post
1
result
editorial
1
result
mini-swe-agent-v2
1
result
swebench-leaderboard
1
result
codesota-shadow-mmlu
1
result
5 of 8 rows marked verified. · first result 2025-12-17, latest 2026-03-15.