BenchLeader

Gemini 3.7 Flash vs Gemini 3.8 Flash

Full profiles: Gemini 3.7 Flash (Google) and Gemini 3.8 Flash (Google): every benchmark, price and rank.

Updated · Data from 12 sources

As of 12 Nov 2026, Gemini 3.7 Flash and Gemini 3.8 Flash are neck and neck on quality (62.0 vs 61.7 on the BenchLeader Index, inside the ±5.2 margin), and they cost about the same at $1.50 vs $1.50 per 1M tokens blended. Choose 3.7 Flash for reasoning (68 vs 63), coding (62 vs 60), and knowledge (66 vs 64) and fast first answers (2.4 s vs 16.8 s); choose 3.8 Flash for agents & tools (51 vs 48) and human preference (69 vs 68). Gemini 3.8 Flash's current price runs until 31 Dec 2026, then $3.00.

Key facts (as of 12 Nov 2026)

  • Gemini 3.7 Flash scores 62.0 on the BenchLeader Index and Gemini 3.8 Flash scores 61.7 (BenchLeader, 12 Nov 2026).
  • Gemini 3.7 Flash costs $1.50 per 1M tokens blended (list price, 12 Nov 2026).
  • Gemini 3.8 Flash costs $1.50 per 1M tokens blended until 31 Dec 2026, then $3.00 (https://ai.google.dev/gemini-api/docs/pricing, checked 20 Oct 2026).
  • Of the 74 benchmarks both models have, Gemini 3.8 Flash leads 47 (BenchLeader, 12 Nov 2026).
  • On Arena WebDev, Gemini 3.7 Flash is rated 1588 and Gemini 3.8 Flash 1583 (LMArena, as published by 12 Nov 2026).

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingGemini 3.7 Flash62 vs 60 category score; Arena WebDev 1588 vs 1583Medium
Agents & tool useGemini 3.8 Flash51 vs 48 category score; Terminal-Bench 18.2% vs 12.4%Medium
MathsEither63 vs 62 category score; FrontierMath Tiers 1–3 71.6% vs 68.4%Either
ReasoningGemini 3.7 Flash68 vs 63 category score; GPQA Diamond 94.8% vs 95.4%Medium
KnowledgeGemini 3.7 Flash66 vs 64 category score; LegalBench 87.3% vs 87.0%Medium
Long documentsEither66 vs 65 category score; AA-LCR v1.1 81.7% vs 81.3%Either
Images and other mediaEither67 vs 68 category score; Arena Vision 1313 vs 1317Either
Chat quality (human preference)Gemini 3.8 Flash69 vs 68 on Arena votesMedium
Lowest costEither$1.50 vs $1.50 per 1M blended; agentic coding session $0.060 vs $0.060; from 1 Jan 2027: $1.50 vs $3.00–
Fastest first answerGemini 3.7 Flash2.4 s (Gemini 3.7 Flash (high)) vs 16.8 s (Gemini 3.8 Flash (high)) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • Gemini 3.7 Flash (high) and Gemini 3.8 Flash (high) are neck and neck on quality (62.0 vs 61.7, within the ±5.2 margin).
  • Gemini 3.7 Flash (high) is stronger in coding, knowledge, long context, maths, reasoning.
  • Gemini 3.8 Flash (high) is stronger in agents & tools, human preference, multimodal.
  • They cost about the same ($1.50 per 1M blended) until 31 Dec 2026, then $1.50 vs $3.00 from 1 Jan 2027.
  • Gemini 3.8 Flash (high) streams 1.3× faster (112 vs 85 tokens per second).

At a glance

SpecGemini 3.7 FlashGemini 3.8 Flash
DeveloperGoogleGoogle
Released13 Aug 20262 Sep 2026
Context window (tokens)1,048,5761,048,576
Compared atGemini 3.7 Flash (high)Gemini 3.8 Flash (high)
Open weightsNoNo
Price in / out / cached input per 1M$0.750 / $3.75 / $0.075$0.750 / $3.75 / $0.075
Announced price changeNone announced$1.50 / $7.50 from 1 Jan 2027

Limits, availability and announced prices from the vendors' own pages: ai.google.dev pricing (checked 20 Oct 2026). Everything else is from the day's data.

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGemini 3.7 Flash (high)Gemini 3.8 Flash (high)Leader
Reasoning67.763.1Gemini 3.7 Flash
Coding62.460.1Gemini 3.7 Flash
Agents & tools47.850.6Gemini 3.8 Flash
Maths62.661.7level
Knowledge65.963.9Gemini 3.7 Flash
Human preference68.269.2Gemini 3.8 Flash
Multimodal67.267.5level
Long context65.665.4level

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75ReasoningCodingAgents & toolsMathsKnowledgeLong contextMultimodalHumanpreferenceGemini 3.7 Flash, Reasoning: 67.7Gemini 3.7 Flash, Coding: 62.4Gemini 3.7 Flash, Agents & tools: 47.8Gemini 3.7 Flash, Maths: 62.6Gemini 3.7 Flash, Knowledge: 65.9Gemini 3.7 Flash, Long context: 65.6Gemini 3.7 Flash, Multimodal: 67.2Gemini 3.7 Flash, Human preference: 68.2Gemini 3.8 Flash, Reasoning: 63.1Gemini 3.8 Flash, Coding: 60.1Gemini 3.8 Flash, Agents & tools: 50.6Gemini 3.8 Flash, Maths: 61.7Gemini 3.8 Flash, Knowledge: 63.9Gemini 3.8 Flash, Long context: 65.4Gemini 3.8 Flash, Multimodal: 67.5Gemini 3.8 Flash, Human preference: 69.2

Not drawn, measured for fewer than two of these models: Instruction following.

Model, click to add or removeIndex

Where sources disagree

Across the 74 benchmarks both have, Gemini 3.7 Flash leads 26, Gemini 3.8 Flash leads 47 and 1 is a tie. Every overall index that rates both puts Gemini 3.8 Flash ahead.

Overall indexGemini 3.7 Flash (high)Gemini 3.8 Flash (high)Leader
BenchLeader Index62.061.7neck and neck
Arena Text14891498Gemini 3.8 Flash

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGemini 3.7 Flash (high)Gemini 3.8 Flash (high)
Blended price per 1M tokens, until 31 Dec 2026$1.50$1.50
Input / output per 1M$0.750 / $3.75$0.750 / $3.75
Blended price per 1M from 1 Jan 2027$1.50$3.00
Input / output per 1M from 1 Jan 2027$0.750 / $3.75$1.50 / $7.50
Cached input per 1M$0.075$0.075
Price per task: chat reply$0.0014$0.0014
Price per task: summarize a 30-page report$0.011$0.011
Price per task: code edit$0.010$0.010
Price per task: agentic coding session$0.060$0.060
Price per task: structured extraction$0.0023$0.0023
Agentic coding session, input mostly cached$0.030$0.030
Cost per benchmark task (AA Index)–$7.77
Output tokens per task (AA Index)–99k
Of which reasoning–60k
Time per task (AA Index)–15 min
Output speed85 tok/s112 tok/s
Time to first answer2.4 s16.8 s
Time to a complete answer–51.5 s

Each model is compared at its best-ranked reasoning-effort setting. Gemini 3.7 Flash (high) scores 62.0 and takes 2 s to its first answer token; Gemini 3.7 Flash (medium) scores 60.1 in 2 s, Gemini 3.7 Flash (low) scores 57.3 in 2 s. Gemini 3.8 Flash (high) scores 61.7 and takes 17 s to its first answer token; Gemini 3.8 Flash (medium) scores 60.3 in 12 s, Gemini 3.8 Flash (low) scores 56.0 in 4 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

153 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGemini 3.7 Flash (high)Gemini 3.8 Flash (high)
BenchLeader IndexBenchLeader62.061.7
Agents & tools scoreBenchLeader47.850.6
Coding scoreBenchLeader62.460.1
Human preference scoreBenchLeader68.269.2
Knowledge scoreBenchLeader65.963.9
Long context scoreBenchLeader65.665.4
Maths scoreBenchLeader62.661.7
Multimodal scoreBenchLeader67.267.5
Reasoning scoreBenchLeader67.763.1
Blended price $/MBenchLeader$1.50$1.50
Output speedBenchLeader85 tok/s112 tok/s
Time to first answerBenchLeader2.4 s16.8 s
Context windowBenchLeader1,048,576 tokens1,048,576 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub94.8%95.4%
Humanity's Last ExamScale AI / CAIS–44.5%†
SimpleBenchSimpleBench–82.4%†
Arena Hard PromptsArena (LMArena)15101520
LiveBench 2.0 ReasoningLiveBench–89.3%†
LiveBench 1.0 ReasoningLiveBench87.8%89.3%
GPQA Diamond (AA)Artificial Analysis94.5%95.3%
Humanity's Last Exam (AA)Artificial Analysis47.9%47.8%
HLE-Diamond (AA)Artificial Analysis–33.1%
GPQA Diamond (Vals)Vals AI93.9%94.4%
ARC-AGI-1ARC Prize95.5%98.5%
ARC-AGI-2ARC Prize84.6%89.2%
ARC-AGI-3ARC Prize–35.0%
CritPtArtificial Analysis14.3%18.3%
CritPt v2 hard subset (AA)Artificial Analysis–46.5%
MysteryMechanismVals AI–36.5%
Chess PuzzlesEpoch AI Benchmarking Hub66.0%58.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub37.0%47.0%
LMCAEpoch AI Benchmarking Hub50.4%51.5%†
DTBenchEpoch AI Benchmarking Hub96.8%95.7%†
ForecastBenchForecastBench66.2%–
Coding
SciCodeSciCode57.2%56.6%
WeirdMLWeirdML–84.8%
FrontierCodeCognition43.6%†41.2%†
Arena CodingArena (LMArena)15211531
Show all 153 benchmarks
MetricPublisherGemini 3.7 Flash (high)Gemini 3.8 Flash (high)
Coding
Arena WebDevArena (LMArena)15881583
LiveBench 2.0 CodingLiveBench–73.3%†
LiveBench 1.0 CodingLiveBench78.9%72.5%
LiveCodeBenchVals AI88.7%89.5%
IOIVals AI67.8%56.9%
SWE-bench (Vals)Vals AI80.8%80.0%
Code MigrationVals AI34.8%36.5%
Code Migration v1.1Vals AI–38.6%
ProgramBenchVals AI0.0%1.0%
SWE-bench MultimodalVals AI–40.8%
Vibe Code Bench 1-100Vals AI–18.8%
Vibe Code Bench v1.1Vals AI70.4%78.7%
SWE Atlas: Codebase QnAScale AI SEAL–47.0%†
SWE Atlas: RefactoringScale AI SEAL–44.8%†
SWE Atlas: Test WritingScale AI SEAL–53.7%†
MirrorCodeEpoch AI Benchmarking Hub41.1%38.9%
Surface Evolver BenchEpoch AI Benchmarking Hub–76.9%
DeepSWE v1.1Epoch AI Benchmarking Hub65.3%73.8%
CursorBenchCursor–39.6%
ALE-BenchEpoch AI Benchmarking Hub904.31269.7
FrontierSWEEpoch AI Benchmarking Hub20.3%19.6%
Terminal-Bench 4.0 (AA)Artificial Analysis13.6%19.7%
Terminal-Bench 2.1 (AA)Artificial Analysis85.8%87.6%
Startrise FrontendStartrise–64.6%†
GameBench 2 OverallDuelLab GameBench 2–58.7†
GameBench 2 by reasoning settingDuelLab GameBench 2–63.7
Design Arena WebsiteDesign Arena1298†1295†
FrontierCode 1.1 (AA)Artificial Analysis–36.4%
FrontierCode 2.0 (AA)Artificial Analysis–18.7%
Agents & tools
Terminal-BenchTerminal-Bench12.4%18.2%
Remote Labor IndexScale AI / CAIS5.0%†5.8%†
APEX-AgentsMercor67.8%†64.3%†
Arena AgentArena (LMArena)-2.22.3
LiveBench Long AgenticLiveBench–16.5%†
LiveBench BrowserLiveBench–45.1%†
LiveBench Agentic CodingLiveBench58.3%54.2%
Terminal-Bench 2.1 (Vals)Vals AI77.5%81.3%
HiL-BenchScale AI SEAL–41.5%†
GDPval-AA v2.1, normalised scoreArtificial Analysis–46.8%
GDPval-AA v2.1Artificial Analysis–1445
τ³-Banking (AA)Artificial Analysis32.8%45.0%
ITBench SRE (AA)Artificial Analysis–52.5%
Analyst Agent (AA)Artificial Analysis60.0%46.3%
CUA-benchVals AI–4.2%
CyberBenchVals AI43.8%43.8%
Finance Agent v2Vals AI59.0%61.4%
Finance Agent v2.1Vals AI–65.8%
Harvey's Legal Agent Benchmark v1.1Vals AI–12.1%
Harvey's Legal Agent Benchmark v1Vals AI8.8%10.0%
Legal Research BenchVals AI34.6%38.9%
SkillsBenchVals AI65.9%58.0%
SREBenchVals AI4.6%11.1%
Tax Agent BenchVals AI22.8%32.1%
Terminal-Bench 4.0 (Vals)Vals AI12.1%19.7%
Terminal-Bench ScienceVals AI7.1%8.6%
Time Horizon Index: KSPVals AI–18.8%
Vending-Bench 2Andon Labs–5093.8†
AutomationBenchArtificial Analysis62.0%59.9%
GDP.pdf (AA)Artificial Analysis23.6%21.0%
Harvey LAB-AA v1.1Harvey–0.0%
EnterpriseOps-GymArtificial Analysis50.4%†–
AA-Briefcase v1.1Artificial Analysis–1203
Design Arena AgenticDesign Arena1216†1218†
Terminal-Bench-Science 0.1 (AA)Artificial Analysis–10.0%
CWE-Bench (AA)Artificial Analysis–52.5%
DeepSec-Bench (AA)Artificial Analysis–14.5%
CyberGym end-to-end (AA)Artificial Analysis–8.4%
FACTS SearchGoogle–80.4%†
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub71.6%68.4%
FrontierMath Tier 4Epoch AI Benchmarking Hub36.6%21.9%
OTIS Mock AIMEEpoch AI Benchmarking Hub97.2%98.9%
ProofBenchVals AI58.0%†48.0%†
LiveBench MathematicsLiveBench93.5%91.6%
Arena MathsArena (LMArena)15021520
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub69.2%69.7%
LiveBench 2.0 Data AnalysisLiveBench–54.0%†
LiveBench ResearchLiveBench–79.4%†
LiveBench 1.0 Data AnalysisLiveBench68.0%54.0%
AA-Omniscience v1.1Artificial Analysis–23.3
MMLU-ProVals AI90.1%90.2%
LegalBenchVals AI87.3%87.0%
TaxEvalVals AI74.7%74.5%
PRBench FinanceScale AI SEAL–49.5%†
PRBench LegalScale AI SEAL–48.9%†
BioMysteryBenchVals AI–62.2%
Excel Modeling BenchmarkVals AI71.3%72.2%
Excel Modeling Benchmark v1.1Vals AI–72.9%
MedCodeVals AI53.4%48.1%
MedScribeVals AI83.9%84.5%
Public Benefits Bench v1.1Vals AI–65.3%
GDP.pdfEpoch AI Benchmarking Hub23.8%23.2%
AA-Omniscience v1.1: accuracyArtificial Analysis–50.5%
AA-Omniscience v1.1: non-hallucinationArtificial Analysis–44.9%
HealthBench Professional (AA)OpenAI–51.7%
Instruction following
LiveBench LanguageLiveBench85.5%87.8%
LiveBench Instruction FollowingLiveBench79.9%81.4%
Arena Instruction FollowingArena (LMArena)14861492
Human preference
Arena TextArena (LMArena)14891498
Arena Creative WritingArena (LMArena)14871489
Arena Multi-turnArena (LMArena)14901505
Arena Longer QueriesArena (LMArena)15011511
Multimodal
Arena VisionArena (LMArena)13131317
LiveBench DesignLiveBench–69.7%†
MMMU-ProArtificial Analysis85.5%85.6%
MMMU-Pro (Vals)Vals AI89.0%89.1%
MortgageTaxVals AI66.7%65.3%
SAGEVals AI–88.2%
AudioMCScale AI SEAL–60.4%
AudioMC (text output)Scale AI SEAL–60.4%
GeoBenchGeoBench–86.0%†
Blueprint-Bench 2Epoch AI Benchmarking Hub–38.6%†
Code2Video-AAArtificial Analysis–899
AA-3DArtificial Analysis–974
Long context
AA-LCR v1.1Artificial Analysis81.7%81.3%
Arena DocumentArena (LMArena)14681469
MLCRArtificial Analysis12.2%10.6%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub158.4†158.8†
LiveBench 2.0LiveBench–57.3%†
LiveBench 1.0LiveBench78.8%75.8%
AA Intelligence Index v5.0Artificial Analysis–30.9
AA Intelligence Index v4.3.2Artificial Analysis39.140.9
AA Progress Index v1Artificial Analysis317.4320.3
Vals IndexVals AI–56.0
Vals Finance IndexVals AI–60.5%
Vals RSI IndexVals AI18.3%19.9%
Vals Science IndexVals AI–35.8%
Vals Router IndexVals AI–56.0%†
AA Cyber IndexArtificial Analysis–25.1

Data as of 2026-11-12. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261110-2acac640a228 (https://benchmarks.duellab.org/releases/gb2-20261110-2acac640a228/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • Gemini 3.8 Flash: list price moves to $1.50 / $7.50 per 1M input / output tokens (announced).
  • Gemini 3.8 Flash (high): 1 new result (MirrorCode).
  • Gemini 3.7 Flash (high): BenchLeader Index 62.6 → 62.1.
  • Gemini 3.7 Flash (high): 1 score revision.
  • Gemini 3.8 Flash (high): 2 score revisions.
  • Gemini 3.7 Flash (high): 1 new result (Arena Document).
  • Gemini 3.8 Flash (high): 1 new result (Arena Document); 1 score revision.
  • Gemini 3.8 Flash (high): 2 score revisions.
  • Gemini 3.8 Flash (high): 1 new result (MirrorCode).
  • Gemini 3.7 Flash (high): 1 new result (MirrorCode).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

Gemini 3.7 Flash vs Gemini 3.8 Flash: questions

Is Gemini 3.7 Flash better than Gemini 3.8 Flash?
Gemini 3.7 Flash (high) and Gemini 3.8 Flash (high) are neck and neck on quality (62.0 vs 61.7, within the ±5.2 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 12 Nov 2026, so check the category scores for your use.
Gemini 3.7 Flash vs Gemini 3.8 Flash: which is cheaper, and will that change?
They cost the same: $1.50 per million tokens blended. Gemini 3.8 Flash's current price runs until 31 Dec 2026, then $3.00.
Is Gemini 3.7 Flash or Gemini 3.8 Flash better for coding?
Gemini 3.7 Flash scores higher in coding (62 vs 60 on the category index, where 50 is average). On Arena WebDev, the top shared coding benchmark, it is 1588 for Gemini 3.7 Flash against 1583 for Gemini 3.8 Flash.
Which is faster, Gemini 3.7 Flash or Gemini 3.8 Flash?
Gemini 3.8 Flash streams faster (112 vs 85 output tokens per second, Artificial Analysis 7-day median).
Is Gemini 3.7 Flash or Gemini 3.8 Flash better for agents and tool use?
Gemini 3.8 Flash scores higher in agentic tasks (51 vs 48 on the category index, where 50 is average). On Terminal-Bench, the top shared agents & tools benchmark, it is 12.4% for Gemini 3.7 Flash against 18.2% for Gemini 3.8 Flash.
More questions (4)
Which is better for long documents?
They are level on long-context benchmarks (66 vs 65). Both accept 1,048,576 tokens of context.
Which has lower latency?
Gemini 3.7 Flash answers first: 2.4 s against 16.8 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on Gemini 3.7 Flash vs Gemini 3.8 Flash?
Across the 74 benchmarks both have, Gemini 3.7 Flash leads 26, Gemini 3.8 Flash leads 47 and 1 is a tie. Every overall index that rates both puts Gemini 3.8 Flash ahead.
When was Gemini 3.8 Flash released?
Gemini 3.8 Flash was released on 2 Sep 2026.