NEWHow AI models draw the world map · Gecko Tests
Beta
Home/Models/Gemini 3.8 Flash
Google DeepMind logo

Gemini 3.8 Flash

by Google DeepMind · Released Sep 2026

Multimodal1M Context
67.7
avg score
Rank #48
Compare
Better than 85% of all models
Context
1.0M tokens (~524 books)
Input $/1M
$0.75
Output $/1M
$3.75
Type
multimodal
License
Proprietary
Benchmarks
25 tested
Data as of
About

Gemini 3.8 Flash is Google's most intelligent Flash model with significant gains from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning.

Tested on 25 benchmarks · BenchGecko score 67.7. Top scores: OTIS Mock AIME 2024-2025 (98.9%), Artificial Analysis · GPQA Diamond (95.3%), GPQA diamond (93.9%).

Looking for similar performance at lower cost?
DeepSeek V4 Flash scores 68.6 (101% as good) at $0.03/1M input · 97% cheaper
Capabilities
coding
44.9
#117 globally
math
63.1
#70 globally
knowledge
66.1
#34 globally
agentic
64.3
#10 globally
general
54.0
#24 globally
speed
58.9
#12 globally
Benchmark Scores
Compare All
Tested on 25 benchmarks · Ranked across 6 categories
Score Distribution (all 312 models)
0255075100
▲ You are here
codingCompare coding →
Deepswe
73.8·
Frontiercode
41.2·
Frontierswe
19.6·
mathCompare math →
OTIS Mock AIME 2024-2025

Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.

98.9·
FrontierMath-Tiers-1-3-v2-Private
68.4·
FrontierMath-Tier-4-v2-Private
21.9·
knowledgeCompare knowledge →
GPQA diamond

Graduate-level science questions written by PhD experts. Diamond subset contains questions where experts disagree, testing deep understanding.

93.9·
SimpleQA Verified

Simple factual questions with verified correct answers. Tests accuracy of basic knowledge retrieval. Low scores indicate hallucination.

69.7·
Chess Puzzles

Tactical chess puzzles testing pattern recognition and multi-move calculation. Measures strategic reasoning ability.

59.0·
Quick compare:xAI logovs Grok 3 BetaOpenAI logovs GPT-5.3-Codexvs PaLM 2-M
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · Google DeepMind Gemini 3.8 Flash
Sep 2026
67.7
$0.75/M in1.0Mctx25 benchmarks
Sep 2026
$0.38/M in(-0.38)1.0Mctx
See the full Gemini 3.8 Flash family →
Recently Happened
Gemini 3.8 Flash added
Sep 3, 2026
Similar Models
xAI logo
Grok 3 Beta
xAI
67.9$3.00/1M
OpenAI logo
GPT-5.3-Codex
OpenAI
68.1$1.75/1M
PaLM 2-M
Unknown
68.1n/a
Links
Info
Google DeepMindPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@Google DeepMind
BenchGecko API
gemini-3-8-flash
Specifications
  • Typemultimodal
  • Context1.0M tokens (~524 books)
  • ReleasedSep 2026
  • LicenseProprietary
  • StatusActive
  • Cost / Message~$0.005
Available On
Google DeepMind logoGoogle DeepMind$0.75
Categories
codingmathknowledgeagenticgeneralspeed
Learn More
context-windowtransformertokens
Share & Export
Tweet
Related Models
xAI logoGrok 3 Beta
67.9
OpenAI logoGPT-5.3-Codex
68.1
PaLM 2-M
68.1
Mistral AI logoMixtral 8x7B Instruct
68.2
Gemini 3.8 Flash is a proprietary multimodal AI model by Google DeepMind, released in September 2026. It has an average benchmark score of 67.7. Context window: 1M tokens.

Related Models

Grok 3 Beta · xAIGPT-5.3-Codex · OpenAIPaLM 2-M · UnknownMixtral 8x7B Instruct · Mistral AIDeepSeek V4 Flash · DeepSeek

Benchmarks

OTIS Mock AIME 2024-2025Artificial Analysis · GPQA DiamondGPQA diamondDtbenchArtificial Analysis · MMMU Pro

Related Pages

Google DeepMind · ProviderGoogle DeepMind · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary

Key facts · as of 2026-10-05

  • Gemini 3.8 Flash by Google DeepMind. BenchGecko score 67.7, rank 48 of 312 scored models (normalized average of public benchmark scores).
  • List price $0.75 input · $3.75 output per 1M tokens (as of 2026-10-05).
  • Sold by 6 providers (as of 2026-10-05): Google $0.38 in / $1.88 out · Google AI Studio $0.38 in / $1.88 out · Google $0.75 in / $3.75 out · Google AI Studio $0.75 in / $3.75 out · Google $1.35 in / $6.75 out · and 1 more. Every provider
  • Gecko Tests · Gecko Score 77 (rank 1): Who Are You B (Knows who made it) · World Map A (97.4% of the map right) · Tokenizer Tax B (44% more tokens outside English).

How to cite · data as of 2026-10-05

Gemini 3.8 Flash · benchmarks, pricing and providers. BenchGecko, data as of 2026-10-05. https://benchgecko.ai/model/gemini-3-8-flash

Credit "Source: BenchGecko" with a link. Prices per provider and Gecko Tests are BenchGecko data (CC BY 4.0); benchmark scores keep their original source, listed in the JSON. JSON · llms.txt · MCP