VerdictPal
Join

Models / Google

Gemini 3 Flash Preview

Gemini 3 · Released 2025-12-17

Google's December 2025 speed tier. 1M context, $0.50/$3 per 1M tokens. Vals Index 49.31 (#9), SWE-bench Verified 78.0%. Pro-level reasoning at Flash pricing.

Default Gemini route when 3.5 Flash is overkill and 3.1 Pro is too expensive. Strong on agentic coding at mid-tier cost.

#14 of 21 on Vals Index · current snapshot
49 Vals Index · 19 benchmark rows · 7 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite57
Context1M
Input / 1M$0.50
Output / 1M$3.0
Knowledge cutoff2025-01-31
Output speed218 t/s
TTFT (default API)0.72s
Statussolid
GeminiNotebookLM
KnowledgeReasoningMathCodingAgenticLong contextTool use
Knowledge
88
Reasoning
81
Math
56
Coding
80
Agentic
49
Long context
51
Tool use
79

69average across 7 tested families

Benchmark placements

Where Gemini 3 Flash Preview places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
MMLU-ProKnowledge#3/ 2188Artificial Analysis2026-07-21
LiveCodeBenchCoding#5/ 1380Artificial Analysis2026-07-21
AA output speedPerformance#8/ 69Artificial Analysis2026-07-21
AIMEMath#14/ 1956Artificial Analysis2026-07-21
τ³-Bench BankingAgentic#17/ 6443Artificial Analysis2026-07-21
SciCodeCoding#29/ 6550Artificial Analysis2026-07-21
IFBenchReasoning#46/ 5755Artificial Analysis2026-07-21
GPQA DiamondReasoning#48/ 6981Artificial Analysis2026-07-21
AA time to first tokenPerformance#51/ 69Artificial Analysis2026-07-21
Artificial Analysis Intelligence IndexReasoning#51/ 6727Artificial Analysis2026-07-21
Humanity's Last ExamReasoning#53/ 6514Artificial Analysis2026-07-21
AA-LCRLong context#55/ 6448Artificial Analysis2026-07-21
Terminal-BenchAgentic#58/ 6932Artificial Analysis2026-07-21
Family context

The three highest-scoring carded models in each capability family. Where Gemini 3 Flash Preview shows up, it's highlighted.

KnowledgeGemini 3 Flash Preview · 88
  1. 1
    GPT 5.5
    91
  2. 2
    Claude Opus 4.8
    91
  3. 3
    Claude Sonnet 4.6
    91
ReasoningGemini 3 Flash Preview · 81
  1. 1
    Claude Mythos Preview
    95
  2. 2
    GPT-5.6 Sol
    94
  3. 3
    Kimi K3
    94
MathGemini 3 Flash Preview · 56
  1. 1
    OpenAI o3
    99
  2. 2
    OpenAI o1
    97
  3. 3
    Gemini 3 Pro
    96
CodingGemini 3 Flash Preview · 80
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticGemini 3 Flash Preview · 49
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextGemini 3 Flash Preview · 51
  1. 1
    Kimi K3
    75
  2. 2
    GPT-5.1
    75
  3. 3
    GPT-5.6 Sol
    74
Tool useGemini 3 Flash Preview · 79
  1. 1
    Claude Sonnet 4.6
    89
  2. 2
    GPT 5.5
    87
  3. 3
    Claude Opus 4.8
    87
Newest receipts

Benchmark rows added to the public ledger for Gemini 3 Flash Preview in the last 120 days. Older rows live in the full table below.

18 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context48%Artificial Analysis2
AA output speedPerformance218 t/sArtificial Analysis2
AA time to first tokenPerformance0.72sArtificial Analysis2
AIMEMath55.67%Artificial Analysis2
Artificial Analysis Intelligence IndexReasoning27.4%Artificial Analysis2
GPQA DiamondReasoning81.2%Artificial Analysis2
Humanity's Last ExamReasoning14.1%Artificial Analysis2
IFBenchReasoning55.1%Artificial Analysis2
LiveCodeBenchCoding79.7%Artificial Analysis2
MMLU-ProKnowledge88.2%Artificial Analysis2
SciCodeCoding49.9%Artificial Analysis2
τ³-Bench BankingAgentic43.27%Artificial Analysis2
Terminal-BenchAgentic31.82%Artificial Analysis2
Vals IndexAgentic49.31%Vals AI — Vals Index49
SWE-bench VerifiedCoding78%SWE-bench official leaderboard54
BFCLTool use79.4%Berkeley Function Calling Leaderboard V499
LiveBenchReasoning63.5%LiveBench leaderboard99
LongBench v2Long context51.4%LongBench v2 leaderboard113
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Knowledge88
Reasoning81
Math56
Coding80
Agentic49
Long context51
Tool use79
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMiniMaxxAIMetaZhipuXiaomiMeituanCohere

Full benchmark ledger

Every catalog benchmark for Gemini 3 Flash Preview. Scores link to the original source; gaps mean no public row exists yet.

19 sourced rows

19 of 32 catalog benchmarks have a sourced row for Gemini 3 Flash Preview.

Knowledge

Reasoning

Math

Coding

Agentic

Long context

Tool use

Performance

Sources