AIEvals

Models

Gemini 3.7 Flash

Specifications and API prices for Gemini 3.7 Flash, recorded by Artificial Analysis.

GoogleReleased 13 August 2026Context 1M tokensOutput Not published in the checked sourcePrivate weights$0.75 in / $3.75 out per Mtok

Family Gemini 3.743 benchmarks with a published result1 independent first place among the 46 models this site listsRead 17 October 2026

Where it stands

Rank compares tracked models with a score. Published shows the source’s rank and field size, when available.

BenchmarkScoreRankPublishedLeaderBehind
Vals IndexVals AI51.27%23of 3324of 41Gemini 4 Argon17.63 pts
Artificial Analysis Intelligence IndexArtificial Analysis3929of 42no list readClaude Opus 5.519 pts
Terminal-Bench 4.0Artificial Analysis13.64%30of 40no list readClaude Sonnet 5.550.00 pts
SWE-bench (Verified)Vals AI80.80%17of 2323of 88Claude Opus 516.20 pts
LiveCodeBenchVals AI88.65%5of 215of 143Claude Fable 5.11.87 pts
GPQA DiamondVals AI93.94%5of 235of 138Gemini 3.1 Pro1.51 pts
MMLU ProVals AI90.12%6of 246of 138Claude Fable 5.12.26 pts
MMMU ProVals AI88.96%5of 205of 93Claude Fable 5.11.68 pts
Vibe Code Bench v1.1Vals AI70.39%29of 3833of 106Claude Sonnet 5.522.00 pts
ProgramBenchVals AI0.00%19of 2923of 54Claude Opus 5.518.50 pts
τ³-BankingArtificial Analysis32.78%25of 30no list readQwen3.8 Max18.56 pts
AA-AnalystAgentArtificial Analysis60.00%1of 18no list readGemini 3.7 Flash—
MLCR-AAArtificial Analysis15.00%31of 37no list readClaude Sonnet 5.560.00 pts
Vals RSI IndexVals AI18.27%18of 2018of 23Claude Opus 5.519.04 pts
Harvey LAB-AAArtificial Analysis90.66%11of 20no list readMuse Spark 1.3 Max4.82 pts
EnterpriseOps-Gym-AAArtificial Analysis50.40%2of 22no list readClaude Fable 50.72 pts
AutomationBench-AAArtificial Analysis62.03%11of 40no list readGemini 4 Argon15.48 pts
SciCodeArtificial Analysis57.18%15of 40no list readClaude Opus 5.59.72 pts
GDP.pdfArtificial Analysis23.60%12of 40no list readGPT-6.1 Sol7.40 pts
CritPtArtificial Analysis14.29%35of 42no list readGPT-5.6 Sol18.00 pts
AA-OmniscienceArtificial Analysis26.4817of 42no list readClaude Opus 5.519.94
AA-LCR v1.1Artificial Analysis81.67%22of 42no list readKimi K37.00 pts
Terminal-Bench 2.1Vals AI77.53%14of 3514of 75Claude Opus 5.510.11 pts
ProofBench v1.1Vals AI58.00%19of 3421of 46Claude Sonnet 5.542.00 pts
LegalBenchVals AI87.26%5of 275of 149Claude Fable 51.30 pts
Harvey's Legal Agent BenchmarkVals AI8.75%15of 3719of 73Muse Spark 1.216.67 pts
SkillsBenchVals AI65.89%3of 203of 35DeepSeek V4.1 Flash3.91 pts
Code MigrationVals AI34.80%29of 3733of 71Claude Sonnet 5.535.03 pts
CyberBench v1.1Vals AI43.75%27of 3138of 43GPT-6 Sol34.23 pts
EMBVals AI71.33%11of 3711of 69Claude Fable 5.15.34 pts
Finance Agent (v2)Vals AI59.04%5of 375of 73Gemini 4 Argon6.36 pts
IOIVals AI67.83%13of 3113of 38Gemini 4 Argon32.17 pts
Legal Research BenchVals AI34.62%31of 3736of 72Muse Spark 1.3 Max20.67 pts
MedCodeVals AI53.39%7of 359of 104Claude Opus 510.18 pts
MedScribeVals AI83.94%27of 3638of 106Claude Opus 5.57.49 pts
SAGEVals AI49.23%12of 3223of 90Claude Opus 4.76.87 pts
SRE BenchVals AI4.58%9of 159of 16GPT-6 Astra52.29 pts
Tax Agent BenchVals AI57.66%31of 3737of 64Claude Fable 5.119.98 pts
DeepSWE v1.1Datacurve65.49%12of 2012of 28GPT-6 Astra8.63 pts
GDPval-AA v2.1Artificial Analysis1,371.2534of 42no list readClaude Opus 5.5474.92
AA-Briefcase v1.1Artificial Analysis1,106.7437of 41no list readClaude Opus 5.5715.11
Humanity's Last ExamArtificial Analysis47.87%14of 42no list readClaude Opus 5.513.48 pts

Nearby on Vals Index

The two models above and below on the Vals Index (Vals AI), on every column where this model has a published number.

ModelVals Index%Artificial Analysis Intelligence IndexindexTerminal-Bench 4.0%SWE-bench%LiveCodeBench%GPQA Diamond%MMLU Pro%MMMU Pro%Vibe Code Bench v1.1%ProgramBench%
GPT-5.6 LunaOpenAI51.69%3711.62%93.00%–91.67%86.04%85.03%77.06%0.00%
DeepSeek V4.1 FlashDeepSeek51.32%3926.77%–––––84.74%–
Gemini 3.7 FlashGoogle51.27%3913.64%80.80%88.65%93.94%90.12%88.96%70.39%0.00%
GPT-6 LunaOpenAI51.22%3712.63%–––––81.65%0.50%
Kimi K3Moonshot AI50.70%4412.63%93.40%87.19%92.93%87.97%88.15%84.96%2.00%
Bold marks a figure better than this model’s on the same column, whichever model it belongs to. Ranks cover the models shown from the stated source.

Cost and speed, as published

Additional measurements from Artificial Analysis’ own row. Read 30 September 2026.

As publishedCost per taskOutput speed
Gemini 3.7 Flash (high)$0.93289.5 tok/s

Results by benchmark

Independent and vendor results retain their own harnesses and effort settings.

BenchmarkIndependent runAs reported by a publisher
τ³-BankingTool use32.78%Artificial Analysisnot published
AA-AnalystAgentData analysis60.00%Artificial Analysisnot published
MLCR-AAMedical reasoning15.00%Artificial Analysisnot published
Vals RSI IndexAutonomous research18.27%Vals AInot published
Harvey LAB-AALegal reasoning90.66%Artificial Analysisnot published
EnterpriseOps-Gym-AAEnterprise workflows50.40%Artificial Analysisnot published
AutomationBench-AAWorkflow automation62.03%Artificial Analysisnot published
SciCodeScientific coding57.18%Artificial Analysisnot published
GDP.pdfDocument reasoning23.60%Artificial Analysisnot published
CritPtScientific reasoning14.29%Artificial Analysisnot published
AA-OmniscienceFactual knowledge26.48Artificial Analysisnot published
AA-LCR v1.1Long-context reasoning81.67%Artificial Analysisnot published
Terminal-Bench 4.0Agentic coding13.64%Artificial Analysis
12.12%Vals AI
not published
Terminal-Bench 2.1Agentic coding77.53%Vals AI
85.77%Artificial Analysis
not published
SWE-bench (Verified)Software engineering80.80%Vals AInot published
LiveCodeBenchCode generation88.65%Vals AInot published
GPQA DiamondScience reasoning93.94%Vals AI
94.55%Artificial Analysis
not published
MMLU ProKnowledge90.12%Vals AInot published
MMMU ProMultimodal reasoning88.96%Vals AI
85.49%Artificial Analysis
not published
Vibe Code Bench v1.1Product coding70.39%Vals AInot published
Terminal-Bench Science 0.1Agentic science
7.14%Vals AI
not published
ProgramBenchAgentic coding0.00%Vals AInot published
ProofBench v1.1Mathematics58.00%Vals AInot published
LegalBenchProfessional knowledge87.26%Vals AInot published
Harvey's Legal Agent BenchmarkProfessional knowledge8.75%Vals AInot published
SkillsBenchAgentic tooling65.89%Vals AInot published
Vals IndexComposite51.27%Vals AInot published
Code MigrationAgentic coding34.80%Vals AInot published
CyberBench v1.1Security43.75%Vals AInot published
EMBFinancial modelling71.33%Vals AInot published
Finance Agent (v2)Financial analysis59.04%Vals AInot published
IOICompetitive programming67.83%Vals AInot published
Legal Research BenchProfessional knowledge34.62%Vals AInot published
MedCodeHealthcare administration53.39%Vals AInot published
MedScribeHealthcare administration83.94%Vals AInot published
SAGEEducation49.23%Vals AInot published
SRE BenchSecurity4.58%Vals AInot published
Tax Agent BenchProfessional knowledge57.66%Vals AInot published
DeepSWE v1.1Software engineering65.49%Datacurvenot published
GDPval-AA v2.1Knowledge work1,371.25Artificial Analysisnot published
AA-Briefcase v1.1Knowledge work1,106.74Artificial Analysisnot published
Humanity's Last ExamMultidisciplinary reasoning47.87%Artificial Analysisnot published
Artificial Analysis Intelligence IndexComposite39Artificial Analysisnot published
Independent column: the suite named under each figure, one harness per suite. Publisher column: each lab’s own table, reproduced on the benchmark’s page. Publisher ranks compare reported scores in the same table row only; ties share a rank and unreported cells are excluded.

Sources

Read on 17 October 2026.