google

Gemini 3.8 Flash

gemini-flash

Google's most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows

ReasoningTool CallingAttachments

Context Window

1.0M

Max Output

66K

Temperature

Yes

Open Weights

No

Knowledge Cutoff

N/A

Released

2026-09-02

Last Updated

2026-09-02

Modalities

Input:TextImageVideoAudioPDF
→
Output:Text

Benchmarks

BenchmarkScoreMetricHarness
DeepSWE73.7pass@1mini-swe-agent
GDPval-AA1545Elo-
Finance Agent61.4pass@1-
Harvey Legal Agent Benchmark10all-pass rate-
Terminal-Bench89.4pass@1Terminus 2
Terminal-Bench19.1pass@1-
GDP.PDF35all-pass rate-
CharXiv Reasoning86.2pass@1-
LVBench87.8pass@1-
LVBench87.1pass@1-
HLE-Verified54.9accuracy-
OSWorld59partial scoreGemini CUA
BioMysteryBench88.8pass@1-
BioMysteryBench56.5pass@1-
LAB-Bench86.2macro-average-

Available Providers (16)

ProviderInput /1MOutput /1MCache Read /1MCache Write /1MReasoningStatus
Eden AIEden AIRp 27.194Rp 135.967Rp 2.720Rp 1.511effort—
CommandCodeCommandCodeRp 27.194Rp 135.967Rp 2.720—effort—
Perplexity AgentPerplexity AgentRp 13.597Rp 67.984Rp 1.360—effort—
Cloudflare AI GatewayCloudflare AI GatewayRp 13.597Rp 67.984Rp 1.360—effort—
Prime InferencePrime InferenceRp 13.597Rp 67.984——effort—
Vercel AI GatewayVercel AI GatewayRp 13.597Rp 67.984Rp 1.360—effort—
Merge GatewayMerge GatewayRp 13.597Rp 67.984Rp 1.360—✓—
OfoxOfoxRp 13.597Rp 67.984Rp 1.360Rp 753effort—
NanoGPTNanoGPTRp 13.597Rp 67.984Rp 1.360Rp 756effort—
AI/ML APIAI/ML APIRp 13.597Rp 67.984Rp 1.360—effort—
Tempr GatewayTempr GatewayRp 13.597Rp 67.984Rp 1.360—effort—
Standard ComputeStandard ComputeRp 13.597Rp 67.984Rp 1.360Rp 756effort—
NEAR AI CloudNEAR AI CloudRp 13.597Rp 67.984Rp 1.360—✓—
PoePoeRp 13.735Rp 68.671Rp 1.375—effort—
Kilo GatewayKilo GatewayRp 13.597Rp 67.984Rp 1.360Rp 756effort—
OpenRouterOpenRouterRp 13.597Rp 67.984Rp 1.360Rp 756effort—