← Back to directory
Google DeepMind
Gemini 2.0 Flash
Released Dec 11, 2024 · knowledge cutoff 2024-08
- Status
- Superseded
- Location
- United States / United Kingdom
- Modality
- Multimodal
- Context window
- 1M
- Max output
- 8K
- Speed
- —
- Price ($/MTok in / out)
- $0.10 / $0.40
- Cost per task
- —
- Open weights
- No
Benchmarks5 of 8 reported
MMLU-Pro
77.6%
GPQA Diamond
62.1%
SWE-bench Verified
51.8%
Humanity's Last Exam
6.6%
LMArena Elo
1356
Terminal-Bench 2.1
—
AIME (latest)
—
ARC-AGI-2
—
Strengths & weaknesses
Strengths
- Fast enough to feel instant inside a user-facing product
- Calls tools and generates images natively rather than through a separate pipeline
- Cheap enough to run across an entire corpus instead of a sample
Weaknesses
- Falls apart on multi-step reasoning — fine for extraction, not for analysis
- Follows complex formatting instructions unreliably
Included as the price-performance outlier of its generation. SWE-bench Verified 51.8 is Google-reported using an agentic code-execution harness.
For developers
API model strings
Not researched
Licence
Proprietary — weights not released
Retirement
No retirement announced
Lineage
No recorded predecessor