Google

Gemini 4 Carbon

Not publicly available. Artificial Analysis measured this model but marks it as not publicly available (checkered pattern in its charts). Its rank still follows its measurements.ReasoningClosed weights
Intelligence
54.2
Rank 10 of 768 measured models
Output speed
76.8 t/s
Rank 260 of 425 measured models
Price/1M
$8
Rank 660 of 726 measured models
Context window
1M
Rank 90 of 1296 measured models
Cost per task
$7.19
Rank 207 of 218 measured models

Gemini 4 Carbon (Low) scores 54.2 on the Artificial Analysis Intelligence Index, ranking 10th of 768 measured models. The closest models on this index are GPT-6 Astra (56.9), GPT-6.1 Sol (54.6), and Claude Sonnet 5.5 (Default Fallback) (54.5). At $8 per million tokens (3:1 input/output blended price), it is cheaper than GPT-6 Astra ($20), more expensive than GPT-6.1 Sol ($4) and Claude Sonnet 5.5 (Default Fallback) ($4). With an output speed of 76.8 t/s, it is faster than GPT-6 Astra (56.5 t/s) and GPT-6.1 Sol (70.6 t/s), slower than Claude Sonnet 5.5 (Default Fallback) (139 t/s).

Not publicly available

Artificial Analysis marks this model as not publicly available (checkered pattern in its charts). Its measurements remain listed for comparison.

Artificial Analysis Indices
Low
Intelligence54.2
Rank 10 of 768 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 & Accounting56.0
Rank 8 of 241 measured models
Strategy & Ops53.0
Rank 31 of 241 measured models
Legal54.0
Rank 14 of 241 measured models
Engineering60.2
Rank 7 of 240 measured models
Economics56.5
Rank 16 of 248 measured models
Standard Benchmarks
Terminal-Bench 4.062.6%
Rank 4 of 240 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