Noometry

Google, proprietary

Gemini 2.5 Flash

Gemini 2.5 Flash by Google ranks 170th of 354 ranked models on the Noometry Index as of October 2026, with a score of 39.3. Its strongest category is long context, where it ranks 17th. API pricing starts at $0.30 per million input tokens and $2.50 per million output tokens, with a 1.05M-token context window.

Last verified

Noometry rank
#170 of 354
Index score
39.3
Evidence
Confirmed 54 results
Provider
Google
Released
April 17, 2025
Weights
Proprietary
Reasoning
Yes
Context window
1.05M
Max output
66K
Input price
$0.30 / M
Output price
$2.50 / M
Blended price
$0.85 / M
Output speed
152 tokens/s Kagi
Value
#103 of 219
Knowledge cutoff
January 2025
Input
text, image, audio, video, pdf

Category scores

Each category score combines every public result we have in that category.

  1. Coding 35.8
  2. Agentic & Tool Use 30.8
  3. Reasoning 18.1
  4. Math 39.9
  5. Knowledge 36.4
  6. Multimodal 41.8
  7. Multilingual 52.3
  8. Instruction Following 75.7
  9. Long Context 47.5
  10. Writing & Preference 53.8
CategoryScoreRankResults
Coding35.8#2204
Agentic & Tool Use30.8#744
Reasoning18.1#2869
Math39.9#983
Knowledge36.4#1686
Multimodal41.8#323
Multilingual52.3#881
Instruction Following75.7#542
Long Context47.5#172
Writing & Preference53.8#1576

Strengths and weaknesses

Categories where Gemini 2.5 Flash places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Long Context47.5+6.6#17 of 296, top 6%
Instruction Following75.7+4.5#54 of 305, top 18%
Multimodal41.8+3.3#32 of 128, top 25%

Weakest categories

CategoryScorevs medianRank
Reasoning18.1−5.5#286 of 350, top 82%
Coding35.8−2.9#220 of 340, top 65%
Knowledge36.4−0.9#168 of 314, top 54%

Closest competitors

The models ranked just above and below Gemini 2.5 Flash. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
DeepSeek-V3#16639.5$0.4173Compare
Grok 4 Fast#16739.4—577Compare
Olmo 3.1 32b Instruct#16839.4——Compare
Granite 4.2 3b#16939.4——Compare
Step 2 16k Exp 202412#17139.2——Compare
Qwen3 32B#17239.2$1.2286Compare
Gemini 2.0 Pro#17339.1——Compare
Molmo 2 8b#17439.1——Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
SWE-bench Verified (bash only)28.7%#32 of 39, top 83%SWE-bench2025-07-26
Aider Polyglot44%Epoch AI
Aider Polyglot47.1%Epoch AI
Aider Polyglot55.1%#19 of 44, top 44%23KEpoch AI
WeirdML41%Epoch AI
WeirdML41%16KEpoch AI
WeirdML41.9%#71 of 119, top 60%16kEpoch AI
LMArena Coding1424#116 of 294, top 40%LMArena2026-10-08
LMArena Coding1400LMArena2026-10-08
ALE-Bench661.88#73 of 105, top 70%Epoch AI

Agentic & Tool Use

BenchmarkScorePositionSettingSourceDate
Terminal-Bench17.1%Epoch AI
Terminal-Bench17.1%#39 of 41, top 96%Epoch AI
Berkeley Function Calling Leaderboard56.2%#12 of 49, top 25%fcBerkeley Function Calling Leaderboard
TheAgentCompany41.1%#2 of 14, top 15%Epoch AI
BALROG33.5%#13 of 35, top 38%Epoch AI
Vending-Bench 2548.84#49 of 60, top 82%Epoch AI

Reasoning

BenchmarkScorePositionSettingSourceDate
ARC-AGI-21.7%Epoch AI
ARC-AGI-22%16KEpoch AI
ARC-AGI-22.2%1KEpoch AI
ARC-AGI-22.5%#65 of 83, top 79%23KEpoch AI
ARC-AGI-22.1%8KEpoch AI
SimpleBench41.2%#52 of 77, top 68%Epoch AI
Kagi LLM Benchmark44.1%Kagi LLM Benchmark
Kagi LLM Benchmark56.8%#48 of 99, top 49%Kagi LLM Benchmark
ARC-AGI-132.3%Epoch AI
ARC-AGI-133.3%#64 of 83, top 78%Epoch AI
ARC-AGI-133.3%16KEpoch AI
ARC-AGI-116%1KEpoch AI
ARC-AGI-132.3%23KEpoch AI
ARC-AGI-125.8%8KEpoch AI
CritPt1.1%#83 of 134, top 62%Epoch AI
EnigmaEval2.7%#29 of 38, top 77%Epoch AI
LMArena Hard Prompts1422#102 of 297, top 35%LMArena2026-10-08
LMArena Hard Prompts1409LMArena2026-10-08
DTBench76.5%#84 of 151, top 56%Epoch AI
LMCA27.5%#82 of 125, top 66%Epoch AI
Epoch Capabilities Index141.53Epoch AI2025-05-20
Epoch Capabilities Index139.96Epoch AI2025-04-17
Epoch Capabilities Index140.8Epoch AI2025-06-17
Epoch Capabilities Index143.03#93 of 213, top 44%Epoch AI2025-09-25
ForecastBench59Epoch AI
ForecastBench60.6#24 of 72, top 34%Epoch AI

Math

BenchmarkScorePositionSettingSourceDate
OTIS Mock AIME 2024-202573.1%#88 of 173, top 51%Epoch AI2025-04-22
OTIS Mock AIME 2024-202570.8%Epoch AI2025-05-20
Omni-MATH38.5%#28 of 57, top 50%HELM Capabilities
LMArena Math1415#109 of 285, top 39%LMArena2026-10-08
LMArena Math1411LMArena2026-10-08
FrontierMath (Feb 2025 set)4.8%#46 of 68, top 68%Epoch AI2025-12-18
FrontierMath Tier 4 (v1)4.2%#31 of 55, top 57%Epoch AI2025-12-18

Knowledge

BenchmarkScorePositionSettingSourceDate
Humanity's Last Exam12.1%#22 of 41, top 54%Epoch AI
Humanity's Last Exam11%Epoch AI
MMLU-Pro63.9%#39 of 58, top 68%HELM Capabilities
Confabulations (lower is better)16.8%#25 of 51, top 50%Lech Mazur benchmarks
Vectara Hallucination Rate (lower is better)7.8%#35 of 96, top 37%Vectara Hallucination Leaderboard
GPQA (HELM)39%#44 of 57, top 78%HELM Capabilities
LMArena Expert1418LMArena2026-10-08
LMArena Expert1426#101 of 273, top 37%LMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1235LMArena2026-10-09
LMArena Vision1253#53 of 122, top 44%LMArena2026-10-09
GeoBench73%Epoch AI
GeoBench76%#8 of 25, top 32%Epoch AI
VPCT38%Epoch AI
VPCT46.2%#9 of 24, top 38%Epoch AI
SpatialViz-Bench36.9%#3 of 8, top 38%Epoch AI

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1396LMArena2026-10-08
LMArena Non-English1409#89 of 297, top 30%LMArena2026-10-08
LMArena Chinese1449LMArena2026-10-08
LMArena Chinese1450#101 of 285, top 36%LMArena2026-10-08
LMArena French1433#88 of 223, top 40%LMArena2026-10-08
LMArena French1432LMArena2026-10-08
LMArena German1417LMArena2026-10-08
LMArena German1418#79 of 231, top 35%LMArena2026-10-08
LMArena Japanese1396LMArena2026-10-08
LMArena Japanese1405#58 of 211, top 28%LMArena2026-10-08
LMArena Korean1385LMArena2026-10-08
LMArena Korean1385#70 of 213, top 33%LMArena2026-10-08
LMArena Russian1415#87 of 283, top 31%LMArena2026-10-08
LMArena Russian1393LMArena2026-10-08
LMArena Spanish1421#88 of 226, top 39%LMArena2026-10-08
LMArena Spanish1405LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
IFEval89.8%#10 of 57, top 18%HELM Capabilities
LMArena Instruction Following1405#95 of 298, top 32%LMArena2026-10-08
LMArena Instruction Following1394LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
Fiction.LiveBench47.2%Epoch AI
Fiction.LiveBench77.8%#12 of 47, top 26%Epoch AI
LMArena Longer Query1419#93 of 291, top 32%LMArena2026-10-08
LMArena Longer Query1406LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1417#102 of 297, top 35%LMArena2026-10-08
LMArena Text1406LMArena2026-10-08
LMArena Creative Writing1387LMArena2026-10-08
LMArena Creative Writing1400#82 of 295, top 28%LMArena2026-10-08
Short-Story Creative Writing76.5%#21 of 39, top 54%Epoch AI
EQ-Bench Creative Writing1137#90 of 115, top 79%EQ-Bench
WildBench81.7%#22 of 57, top 39%HELM Capabilities
LMArena Multi-Turn1397LMArena2026-10-08
LMArena Multi-Turn1408#113 of 295, top 39%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
google$0.30$2.50$0.032026-10-10
openrouter$0.30$2.50$0.032026-10-10
vertex$0.30$2.50$0.032026-10-10

All Google API prices → Estimate your cost →

Compare Gemini 2.5 Flash

Other Google models

Frequently asked questions

How good is Gemini 2.5 Flash?

Gemini 2.5 Flash by Google ranks 170th of 354 ranked models on the Noometry Index as of October 2026, with a score of 39.3. Its strongest category is long context, where it ranks 17th. API pricing starts at $0.30 per million input tokens and $2.50 per million output tokens, with a 1.05M-token context window.

How much does Gemini 2.5 Flash cost?

Gemini 2.5 Flash costs $0.30 per million input tokens and $2.50 per million output tokens on Google's own API, with cached input at $0.03.

What is Gemini 2.5 Flash's context window?

Gemini 2.5 Flash accepts up to 1.05M tokens of input and can write up to 66K tokens in one response.

Is Gemini 2.5 Flash open source?

No. Gemini 2.5 Flash is proprietary and available only through Google's API and partner platforms.

How fast is Gemini 2.5 Flash?

Gemini 2.5 Flash generated about 152 output tokens per second in the Kagi LLM Benchmark's timed runs. Speed varies by provider, load and reasoning effort.

What are Gemini 2.5 Flash's strengths and weaknesses?

Relative to other ranked models, Gemini 2.5 Flash places best in long context, instruction following, multimodal and lowest in reasoning, coding, knowledge.

What is Gemini 2.5 Flash best at?

Its best category is long context, where it ranks 17th on Noometry.