vs
0:0
overall capability wins (2 ties across 2 scored). Gemini 3 Pro leads overall.
Verdict
Gemini 3.5 Flash Cyber and Gemini 3 Pro are closely matched on the overlapping metrics we have.
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Gemini 3.5 Flash Cyber
- -
- Gemini 3 Pro
- $5
Price per 1M tokens
Full comparison
Organization
Family
Gemini
Gemini
License
Proprietary
Proprietary
Open weights
No
No
Release
Jul 21, 2026
Nov 18, 2025
Knowledge cutoff
-
-
API / provider
No primary API price listed
Modalities
text → text
text, image, audio, video → text
Specs
Context window
1M
1M
Max output
66K
66K
Parameters
—
—
Pricing
Input $/1M
—
$2
Output $/1M
—
$12
Blended $/1M (3∶1)
-
$4.50
Speed
tok/s
-
120
TTFT (s)
-
0.45
Reasoning
MMLU-Pro
—
—
GPQA Diamond
—
91.9
Humanity's Last Exam
—
37.5
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
—
AA-LCR v1.1
—
—
CritPt
—
—
MMMU-Pro
—
—
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
69.6
SWE-bench Pro
—
—
SWE-bench Multilingual
—
68.7
LiveCodeBench
—
—
Terminal-Bench 2.1
—
73.9
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
—
CursorBench
—
—
SWE-Rebench
—
50.7
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
1,439
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
1,486
| Metric | Gemini 3.5 Flash Cyber | Gemini 3 Pro | Delta |
|---|---|---|---|
| Identity | |||
| Organization | — | ||
| Family | Gemini | Gemini | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | Jul 21, 2026 | Nov 18, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | No primary API price listed | — | |
| Modalities | text → text | text, image, audio, video → text | — |
| Specs | |||
| Context window | 1M | 1M | tie |
| Max output | 66K | 66K | tie |
| Parameters | — | — | — |
| Pricing | |||
| Input $/1M | — | $2 | — |
| Output $/1M | — | $12 | — |
| Blended $/1M (3∶1) | - | $4.50 | — |
| Speed | |||
| tok/s | - | 120 | — |
| TTFT (s) | - | 0.45 | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | — | 91.9 | — |
| Humanity's Last Exam | — | 37.5 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | — | — |
| AA-LCR v1.1 | — | — | — |
| CritPt | — | — | — |
| MMMU-Pro | — | — | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | 69.6 | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | 68.7 | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | 73.9 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | 50.7 | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | 1,439 | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | 1,486 | — |
Gemini 3.5 Flash Cyber
- Context
- 1M / 66K out
- Parameters
- —
- Price
- —
- Speed
- —
- Modalities
- text → text
- License
- Proprietary
Cybersecurity-specialized Gemini 3.5 Flash fine-tune for CodeMender — finds, validates, and patches vulnerabilities; limited pilot for governments and trusted partners only (no public API).
Gemini 3 Pro
- Context
- 1M / 66K out
- Parameters
- —
- Price
- $2 / $12
- Speed
- 120 tok/s · 0.45s TTFT
- Modalities
- text, image, audio, video → text
- License
- Proprietary
Gemini 3 Pro preview (Nov 2025); superseded by Gemini 3.1 Pro but still listed in deprecation docs.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.