Gemini 3 Flash

Gemini 3 Flash is a non-reasoning model from Google. 22 benchmarks count toward its score, in 7 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.55.3 ±3.9
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.213/s
Input / 1MUS dollars per 1M input tokens.$0.5
Output / 1MUS dollars per 1M output tokens.$3
ContextMaximum tokens in one request.1M
EloLMArena rating and rank.1467 (#30)

50 is the middle of the board. The range shows the doubt in the index.

30,225 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category. 50 is the middle.

50 is the middle
AgenticMulti-step tasks with tools.
48.6
CodingCode writing and repair.
58.0
ReasoningLogic problems and puzzles.
54.6
MultimodalTasks with images and text.
63.1
KnowledgeFacts and expert knowledge.
58.7
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
45.0
MathMath problems.
58.6

Results

22 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result on the index scale.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
OTIS Mock AIME 2024-2025Math95.6%65.0high effortEpoch AI
Artificial Analysis Global-MMLU-LiteMultilingual92.7%Artificial Analysis
τ²-bench TelecomAgentic91.2%64.3high effort · Sierra2 Mar 2026Sierra Research
GPQA diamondKnowledge89.4%60.8high effortEpoch AI
τ²-bench AirlineAgentic82.5%58.0high effort · Sierra2 Mar 2026Sierra Research
Artificial Analysis GPQA DiamondKnowledge81.2%51.8Artificial Analysis
Artificial Analysis MMMU-ProMultimodal78.6%63.1Artificial Analysis
τ²-bench RetailAgentic76.8%53.9high effort · Sierra30 Apr 2026Sierra Research
SWE-Bench verifiedCoding75.4%58.0Epoch AI
SWE-bench MultilingualMultilingual72.7%mini-SWE-agent20 Feb 2026SWE-bench team
SWE-bench MultilingualCoding72.7%mini-SWE-agent2 Sept 2026SWE-bench team
SimpleQA VerifiedKnowledge66.8%83.4high effortEpoch AI
Gert Labs Composite Game BenchmarkAgentic56.6%62.7Gert Labs
Artificial Analysis Long Context ReasoningReasoning55.3%46.5Artificial Analysis
Artificial Analysis IFBenchInstruction55.1%45.0Artificial Analysis
FrontierMath-Tiers-1-3-v2-PrivateMath51.2%60.3Epoch AI
Claw-EvalAgentic49.2%35.8Bowen Ye et al.
Artificial Analysis Omniscience AccuracyKnowledge45.8%70.5Artificial Analysis
τ²-Bench Tool-Agent-User EvaluationAgentic43.3%29.9Victor Barres et al.
Chess PuzzlesReasoning40.0%75.2high effortEpoch AI
FrontierMath-2025-02-28-PrivateMath35.6%63.9Epoch AI
τ²-bench BankingAgentic27.3%18.4high effort · Sierra4 Aug 2026Sierra Research
Mystery Game PuzzlesReasoning20.0%55.2high effortEpoch AI
Artificial Analysis Intelligence IndexKnowledge17.9%44.8Artificial Analysis
FrontierMath-Tier-4-v2-PrivateMath17.1%56.1Epoch AI
Artificial Analysis Humanity's Last ExamKnowledge15.0%40.8Artificial Analysis
JobBenchAgentic11.4%42.0Yuetai Li et al.
FrontierMath-Tier-4-2025-07-01-PrivateMath4.2%48.0Epoch AI
Critical Physics TasksReasoning1.4%41.3Artificial Analysis

22 benchmarks count, from 26 of 29 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsEpoch AI, collected directlyCC BY — free to use and redistribute with attributionSierra Research, collected directlyMIT — results are in the licensed repositorySWE-bench team, collected directlyNo licence stated for the leaderboard. The harness repo is MIT

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeApodex 1.1 Mini59.1 · Free

More from Google

Gemini 3.7 Flash72.3Gemini 3.8 Flash72.1Gemini 3.5 Flash67.3Gemini 3.6 Flash66.8Gemini 3.1 Pro64.7