Gemini 3.7 Flash

Lab: google-deepmind
Weights: closed
Released: 2026-08-13
Context window: 1,000,000
Modalities: text, image, audio
API: yes
Local-runnable: no
Status: current

Scores

Human preference

Value	Source	As of	Reported by	Notes
1490 elo	lmarena-text ↗	2026-08-31	aggregatorA	gemini-3.7-flash-high, #9; unchanged in value from the 2026-08-26 fetch, as-of refreshed.
1490 elo	lmarena-text ↗	2026-08-26	aggregatorA	gemini-3.7-flash-high variant on the 2026-08-26 board fetch; introductory pricing through 2026-12-31 per Google's announcement.

Events

Independent test finds current AI models unreliable at identifying dangerous mushrooms

The Register reported on September 2, 2026 on an independent test by Piotr Migdal running 1,040 mushroom photos across 16 models. Gemini 3.8 Flash scored best (65% correct on first guess, 85% within its top five), while Qwen3.8-27b scored worst (13% first-guess, and misidentified poisonous mushrooms as edible 36% of the time). Meta's Muse Spark 1.2 had the lowest false-positive rate (8%) largely because it declined to guess in ambiguous cases.

2026-09-02 Safety / alignment Benchmark result AI-assisted mushroom hunting is a recipe for a bad trip

Google ships Gemini 3.8 Flash, its 'most intelligent' Flash-tier model, for long-horizon agents

On September 2, 2026, Google released gemini-3.8-flash as generally available, describing it in the Gemini API changelog as its most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows. No benchmark scores were included in the changelog entry.

2026-09-02 Release Gemini 3.8 Flash generally available (GA)

Gemini adds agentic video understanding, cutting long-video token use up to 88%

On September 1, 2026, Google DeepMind and the Gemini API released agentic video understanding for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite, letting the model dynamically request transcripts, frames, or audio tracks from a video timeline on demand rather than processing it statically. Google reports this uses up to 88% fewer tokens for long-form video content compared to static processing.

2026-09-01 Release Tooling / agents Agentic video understanding

← Back to Models