Inference
EN
Google

Gemini 2.0 Flash

Gemini 2.0 Flash is a standard model from Google. 15 benchmarks count toward its index, in 7 categories.

  • Closed weights
  • Standard
  • Input: text
  • Released 2025-02
RankedEnough results for an index.
IndexAll categories, out of 100.29.5±4.7Median 51.0
CoverageIndex weight that has results.95%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1422

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.1 counted22.4—
CodingCode writing and repair.1 counted11.2—
ReasoningLogic problems and puzzles.3 counted35.1191 / 227
MultimodalTasks with images and text.1 counted37.0—
KnowledgeFacts and expert knowledge.4 counted35.7221 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted26.6—
MathMath problems.4 counted28.1148 / 168

Results

15 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
MATH level 5Math82.2% ±0.926 / 88
Measuring Short-Form Factuality in Large Language ModelsKnowledge29.9%6 / 14
GPQA diamond · Epoch AIKnowledge64.1% ±2.7125 / 217
OTIS Mock AIME 2024-2025Math31.1% ±6.3132 / 200
Massive Multi-discipline Multimodal UnderstandingMultimodal69.3%19 / 28
Artificial Analysis Omniscience AccuracyKnowledge23.5%112 / 164
Critical Physics TasksReasoning0.0%150 / 214
Artificial Analysis IFBenchInstruction40.2%103 / 146
Graduate-Level Google-Proof Q&AKnowledge65.2%103 / 141
τ²-Bench Tool-Agent-User EvaluationAgentic29.5%113 / 149
Artificial Analysis GPQA DiamondKnowledge62.3%153 / 199
ARC-AGI-2 (semi-private)Reasoning1.3%103 / 131
FrontierMath-2025-02-28-PrivateMath1.7% ±0.855 / 69
Artificial Analysis Long Context ReasoningReasoning31.3%179 / 217
SWE-bench Verified · swebench.comCoding13.5%39 / 46
Artificial Analysis Humanity's Last ExamKnowledge4.3%190 / 222
American Invitational Mathematics Examination 2025Math29.7%49 / 53
SWE-bench Verified · SWE-bench/experimentsCoding13.5%61 / 65
Software Engineering Benchmark VerifiedCoding21.4%112 / 114
FrontierMath-2025-02-28-PublicMath0.0%33 / 64

Sources

BenchLMCC BY-NC 4.0Epoch AICC BYSWE-bench teamNo licence statedARC Prize FoundationTerms restrict reuse

History

5 releases
Index30.2 → 29.5
23 Oct 202612 Nov 2026

Questions

How good is Gemini 2.0 Flash?
Its index is 29.5 out of 100. That is place 251 of 283 ranked models.
What is Gemini 2.0 Flash best at?
Knowledge: 35.7 out of 100, place 221 of 268.
What is the context window of Gemini 2.0 Flash?
1M tokens.
Is Gemini 2.0 Flash open source?
No. Its weights are not published.
When was Gemini 2.0 Flash released?
February 2025.

Badge

Gemini 2.0 Flash on Inference360: place 251 of 283, index 29.5
HTML
<a href="https://inference360.ai/models/gemini-2-0-flash"><img src="https://inference360.ai/badges/gemini-2-0-flash.svg" alt="Gemini 2.0 Flash on Inference360: place 251 of 283, index 29.5" height="20"></a>
Markdown
[![Gemini 2.0 Flash on Inference360: place 251 of 283, index 29.5](https://inference360.ai/badges/gemini-2-0-flash.svg)](https://inference360.ai/models/gemini-2-0-flash)

Compare with

LFM2.5-2.6B29.3Grok 229.1DeepSeek V329.9Sarvam 30B30.3Gemini 1.5 Flash20.2

More from Google

All models
Gemini 2.0 Flash-LiteUnrankedGemini 2.5 Pro46.6Gemma 3 27B25.3Gemini 2.5 Pro Preview 06-05UnrankedGemini 2.5 Flash35.0