Back

Google

Gemini 3.8 Flash

ReasoningClosed weights
Intelligence
39.8
Rank 66 of 749 measured models
Coding
74.1
Rank 35 of 298 measured models
Price/1M
$1.50
Rank 436 of 671 measured models
Context window
1M
Rank 86 of 1216 measured models
Cost per task
$0.931
Rank 130 of 198 measured models

Gemini 3.8 Flash (Medium) scores 39.8 on the Artificial Analysis Intelligence Index, ranking 66th of 749 measured models. The closest models on this index are Qwen3.8 2.4T A95B (39.9), Qwen3.8-Flash-Next (39.8), and Muse Spark 1.2 (Xhigh) (39.6). At $1.50 per million tokens (3:1 input/output blended price), it is cheaper than Qwen3.8 2.4T A95B ($3) and Muse Spark 1.2 (Xhigh) ($2), more expensive than Qwen3.8-Flash-Next ($0.23).

Artificial Analysis Indices
Medium
Intelligence39.8
Rank 66 of 749 measured models
Coding74.1
Rank 35 of 298 measured models
Agentic39.7
Rank 47 of 190 measured models
Reasoning levels
Each point is a reasoning level of this model: more reasoning often raises the score but costs more. Select a point to open that level.
Intelligence Indexmost efficient ↗Cost per task
Capability indexes
Performance in specific industries, measured by Artificial Analysis
Finance & Accounting44.7
Rank 55 of 224 measured models
Strategy & Ops48.5
Rank 47 of 224 measured models
Legal51.0
Rank 36 of 224 measured models
Engineering40.5
Rank 63 of 223 measured models
Economics50.3
Rank 57 of 231 measured models
Standard Benchmarks
GPQA Diamond93.5%
Rank 14 of 620 measured models
HLE42.1%
Rank 73 of 669 measured models
SciCode55.1%
Rank 52 of 236 measured models
AA-LCR84.0%
Rank 11 of 580 measured models
Terminal-Bench 2.183.9%
Rank 34 of 245 measured models
Terminal-Bench v419.7%
Rank 62 of 223 measured models
τ-Bench Banking45.8%
Rank 14 of 211 measured models
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy