Gemini 3.8 Flash vs Gemini 4 Argon

Gemini 4 Argon leads on the Visual Intelligence Index by 3.5 points, and Gemini 3.8 Flash scores higher on 2 of the 7 benchmarks where the two differ.

Model

Compared with

GoogleGemini 3.8 Flash80.6%Rank #3 GoogleGemini 4 Argon84.1%Rank #1

At a glance

Measure	Gemini 3.8 Flash	Gemini 4 Argon
Visual Intelligence Index	80.6%	84.1%
Rank	#3 of 10	#1 of 10
Cost per taskMean over the benchmarks, at list price.	$0.0409	$0.110
Latency per taskEach benchmark's median, averaged.	22.7s	23.9s
Tokens per task	52,254	54,614
Input / 1M tokens	$0.75	$2.00
Output / 1M tokens	$3.75	$10.00
Context window	1.05M	1.05M
Availability	Proprietary	Proprietary

How cost and latency are measured

Capabilities

Each capability index is a weighted mean of the benchmarks that contribute to the capability.

Capability	Gemini 3.8 Flash	Gemini 4 Argon	Lead
Perception	79.8%	81.7%	−1.9
Spatial Reasoning	82.5%	85.4%	−2.9
Temporal Reasoning	81.3%	84.5%	−3.1
Action / Event Understanding	78.2%	84.7%	−6.5
Causal Reasoning	81.6%	85.4%	−3.8
Information Retrieval	79.4%	82.7%	−3.3

Benchmarks

Benchmark	Gemini 3.8 Flash	Gemini 4 Argon	Lead
Video-MME v2	71.3%	77.0%	−5.8
LVBench	86.4%	86.3%	+0.1
Perception Test	85.5%	89.7%	−4.2
NExT-QA	88.0%	90.0%	−2.1
Q-Bench Video	69.3%	66.9%	+2.4
EgoSchema	80.8%	86.4%	−5.6
UCF101-AD	63.5%	81.8%	−18.4

Usability

Gemini 3.8 Flash

4.0/ 5

Reachable1.0: There is an endpoint you can call without hosting anything.
First-party API — supported
Third-party API — supported
Portable0.0: You can run it yourself, and are not tied to one vendor.
Open weights — not supported
Self-hostable — not supported
Multimodal input1.0: It takes the footage directly, rather than frames you extracted.
Video ingestion — supported
Audio — supported
Programmable1.0: Output you can parse, and tools it can call on its own.
Structured output — supported
Tool calling — supported
Operable1.0: Usable interactively and in bulk, not only one call at a time.
Streaming — supported
Batch API — supported

Gemini 4 Argon

4.0/ 5

Reachable1.0: There is an endpoint you can call without hosting anything.
First-party API — supported
Third-party API — supported
Portable0.0: You can run it yourself, and are not tied to one vendor.
Open weights — not supported
Self-hostable — not supported
Multimodal input1.0: It takes the footage directly, rather than frames you extracted.
Video ingestion — supported
Audio — supported
Programmable1.0: Output you can parse, and tools it can call on its own.
Structured output — supported
Tool calling — supported
Operable1.0: Usable interactively and in bulk, not only one call at a time.
Streaming — supported
Batch API — supported