Enter opens the highlighted result. Esc closes.

Google Gemini 4 Neon

  • Google
  • Released November 18, 2026
  • No public weights listed
  • 1.05M context

As of November 22, 2026, Google Gemini 4 Neon ranks #2 of 120 models overall with 89.1 ±2.0 on the LLMBenchmark Index. Gemini 4 Neon ranks #3 of 110 in math, its strongest category. Google Gemini 4 Neon costs $3.00 per million input tokens and $15.00 per million output tokens. Gemini 4 Neon was released on November 18, 2026.

Gemini 4 Neon's output price ranks 310 of 346 priced models, counting from the cheapest.

  • Coding index 86.9 Verified, 4 benchmarks
  • Cost per solved task Not computable No published attempt cost
  • Output, per million tokens $15.00 310 of 346 from the cheapest Maker's API price
  • Input, per million tokens $3.00 Maker's API price
  • Context window 1.05M Tokens per request
New model

Gemini 4 Neon launch tracker

Listed on OpenRouter November 18, 2026, 4 days before our data. Results so far on 69 of 222 public boards; this page updates every day as boards publish.

About and specs

Gemini 4 Neon is a fast, mid-priced model in the Gemini 4 family from Google, built for coding and agentic work. It is suited for repository and terminal coding, web...

API model id google/gemini-4-neon
  • Tool callingListed
  • Structured outputsListed
  • JSON modeListed
  • Reasoning modeListed
  • Web searchNot listed
Takes in
Audio, Files, Images, Text, Video
Gives back
Text
Context window
1.05M tokens
Max output
944K tokens
Knowledge cutoff
Not published
Cached input $/Mtok
$0.15
Cache write $/Mtok
$0.0833
Reasoning $/Mtok
$15.00
Web search, per request
$0.014
Tokenizer
Gemini
Requests moderated
No
Withdrawn on
No date announced
Parameters
Not published
Downloads, all time
No public weights
Training compute
No estimate published
Output speed
106 tokens/s, xHigh effort
Speed on OpenRouter
48 tokens/s via Google AI Studio, 30-minute median
First token on OpenRouter
5.2 s via Google AI Studio, 30-minute median

Benchmarks: 69 of 222 measured

The tick on a coding bar is the median of every model measured on it. Outside coding, the place is this model's rank on that board. A board that never ran the model says so.

Coding index

86.9
  • Terminal-Bench 4 (Vals AI) 61.6% (Mini-SWE-agent) 82 of 100
  • Vibe Code Bench (Vals AI) 93.7% (OpenHands) 94 of 100
  • LiveBench coding 89.7% 90 of 100
  • SWE-bench Verified (Vals AI) Not measured
  • FrontierCode Not measured
  • LMArena WebDev Elo, scaled from 1200 to 1900 1765 Elo (High) 81 of 100

The index averages only the coding benchmarks that have a result; the others are skipped, never filled in. How it is weighted

Other public boards

All rankings

Coding8 of 46 boards

2 more coding boards

Not measured on Aider polyglot, SWE-bench Verified, SWE-bench Lite, SWE-bench (full test set), SWE-bench Multimodal, SWE-bench Multilingual, AA Coding Index, LiveCodeBench (Vals AI), Code migration (Vals AI), ProgramBench (Vals AI), IOI (Vals AI), Terminal-Bench 2.1 (Vals AI), SWE-bench Multimodal (Vals AI), Vibe Code Bench 1-100 (Vals AI), SkillsBench (Vals AI), CyberBench (Vals AI), LiveCodeBench (Artificial Analysis run), SciCode, Terminal-Bench Hard (Artificial Analysis run), Terminal-Bench, DeepSWE, FrontierSWE, GSO-Bench, WeirdML, MirrorCode, Cybench, ALE-Bench, LMArena WebDev: from an image, SciCode, SWE-Bench Pro V2, SWE Atlas: Codebase QnA, SWE Atlas: Refactoring, SWE Atlas: Test Writing, Terminal-Bench 2.0, Terminal-Bench 2.1, Terminal-Bench 3.0, Terminal-Bench 4.0, ProgramBench

Intelligence1 of 2 boards

Not measured on Epoch Capabilities Index

Agentic work11 of 44 boards

5 more agentic work boards

Not measured on AA Agentic Index, Design Arena: agents building web apps, Design Arena: agents building full-stack apps, Design Arena: agents building mobile apps, Design Arena: agents building Android apps, Design Arena: agents building Godot games, Design Arena: agents building games, Design Arena: HTML slides, Design Arena: PowerPoint slides, Design Arena: spreadsheets, SREBench (Vals AI), Finance Agent (Vals AI), Excel Modeling Benchmark (Vals AI), Time Horizon Index (Vals AI), CUA-bench (Vals AI), Robo-bench (Vals AI), τ²-Bench Telecom, METR time horizons, Remote Labor Index, GDPval, OSWorld 2.0, BALROG, DeepResearch Bench, PostTrainBench, TheAgentCompany, OSWorld, SpireClimb, LMArena Search, Berkeley Function Calling Leaderboard (BFCL), BrowseComp-Plus, MCP Atlas, Toolathlon-Verified, τ³-Banking

Reasoning12 of 26 boards

6 more reasoning boards

Not measured on GPQA Diamond, MMMU Pro (Vals AI), GPQA Diamond (Artificial Analysis run), Humanity's Last Exam (Artificial Analysis run), ARC-AGI, ARC-AGI-2, Humanity's Last Exam, VPCT, BIG-Bench Hard, LMArena Vision: diagrams, LMArena Vision: homework, Kagi LLM Benchmark, Humanity's Last Exam (Scale run), GoBench

Math5 of 14 boards

Not measured on AIME, MATH 500 (Vals AI), MGSM (Vals AI), AA Math Index, AIME 2025 (Artificial Analysis run), ProofBench, MATH level 5, GSM8K, MathArena

Knowledge6 of 21 boards

Not measured on MMLU Pro, LegalBench (Vals AI), CaseLaw (Vals AI), CorpFin (Vals AI), TaxEval (Vals AI), MortgageTax (Vals AI), MedQA (Vals AI), MedCode (Vals AI), MMLU Pro (Artificial Analysis run), MMLU, GeoBench, AI2 Reasoning Challenge, FACTS Benchmark Suite, PRBench Finance, PRBench Legal

Web design6 of 9 boards

Not measured on Design Arena: data viz, Design Arena: SVG, Design Arena: ASCII art

General16 of 49 boards

10 more general boards

Not measured on Public Benefits Bench (Vals AI), SAGE (Vals AI), MedScribe (Vals AI), Vals Multimodal Index, IFBench, Long context reasoning (AA-LCR), GDPval-AA, DTBench, LMCA, Fiction.LiveBench, Creative writing (Lech Mazur), CL-bench, Surface Evolver Bench, ExploitBench, HellaSwag, WinoGrande, PIQA, CL-bench Life, LMArena Vision, LMArena Text: German, LMArena Text: Japanese, LMArena Text: Korean, LMArena Text: Polish, LMArena Vision: English, LMArena Vision: Chinese, LMArena Vision: reading text, LMArena Vision: humor, LMArena Vision: captioning, LMArena Vision: recognizing things, LMArena Vision: creative writing, LMArena Document, MultiChallenge, EQ-Bench 4

Image generation0 of 5 boards

Not measured on Design Arena: images, Design Arena: graphic design, Design Arena: logos, LMArena Text to Image, LMArena Image Edit

Public boards shown as published (Vals AI, Artificial Analysis, Design Arena) and kept out of the coding index. Each links to its source on the benchmark page.

All 222 benchmarks as a table
Benchmark results for Google Gemini 4 Neon
BenchmarkResultIndex weight
Aider polyglotCoding Not measured Not in index
SWE-bench VerifiedCoding Not measured Not in index
SWE-bench LiteCoding Not measured Not in index
SWE-bench (full test set)Coding Not measured Not in index
SWE-bench MultimodalCoding Not measured Not in index
SWE-bench MultilingualCoding Not measured Not in index
LiveBench codingCoding 89.7% 15%
LMArena WebDevCoding 1765(High) 10%
Vals IndexGeneral 70.9% Not in index
GPQA DiamondReasoning Not measured Not in index
AIMEMath Not measured Not in index
MMLU ProKnowledge Not measured Not in index
AA Intelligence IndexIntelligence 61.4 Not in index
AA Coding IndexCoding Not measured Not in index
AA Agentic IndexAgentic work Not measured Not in index
Design Arena: websitesWeb design 1347 Not in index
Design Arena: UI componentsWeb design 1349 Not in index
Design Arena: game devWeb design 1394 Not in index
Design Arena: data vizWeb design Not measured Not in index
Design Arena: 3DWeb design 1490 Not in index
Design Arena: all code categoriesWeb design 1391 Not in index
Design Arena: SVGWeb design Not measured Not in index
Design Arena: ASCII artWeb design Not measured Not in index
Design Arena: agents building web appsAgentic work Not measured Not in index
Design Arena: agents building full-stack appsAgentic work Not measured Not in index
Design Arena: agents building mobile appsAgentic work Not measured Not in index
Design Arena: agents building Android appsAgentic work Not measured Not in index
Design Arena: agents building Godot gamesAgentic work Not measured Not in index
Design Arena: agents building gamesAgentic work Not measured Not in index
Design Arena: HTML slidesAgentic work Not measured Not in index
Design Arena: PowerPoint slidesAgentic work Not measured Not in index
Design Arena: spreadsheetsAgentic work Not measured Not in index
Design Arena: imagesImage generation Not measured Not in index
Design Arena: graphic designImage generation Not measured Not in index
Design Arena: logosImage generation Not measured Not in index
LiveBench reasoningReasoning 89.6% Not in index
LiveBench data analysisGeneral 85.3% Not in index
LiveBench researchAgentic work 94.5% Not in index
LiveBench browserAgentic work 95.6% Not in index
LiveBench long agenticAgentic work 82.6% Not in index
LiveBench designWeb design 81.5% Not in index
LiveCodeBench (Vals AI)Coding Not measured Not in index
SWE-bench Verified (Vals AI)Coding Not measured 15%
Terminal-Bench 4 (Vals AI)Coding 61.6%(Mini-SWE-agent) 25%
Vibe Code Bench (Vals AI)Coding 93.7%(OpenHands) 20%
Code migration (Vals AI)Coding Not measured Not in index
ProgramBench (Vals AI)Coding Not measured Not in index
SREBench (Vals AI)Agentic work Not measured Not in index
IOI (Vals AI)Coding Not measured Not in index
Terminal-Bench 2.1 (Vals AI)Coding Not measured Not in index
SWE-bench Multimodal (Vals AI)Coding Not measured Not in index
Vibe Code Bench 1-100 (Vals AI)Coding Not measured Not in index
SkillsBench (Vals AI)Coding Not measured Not in index
CyberBench (Vals AI)Coding Not measured Not in index
Finance Agent (Vals AI)Agentic work Not measured Not in index
Tax Agent Bench (Vals AI)Agentic work 54.9% Not in index
Legal Research Bench (Vals AI)Agentic work 56.3% Not in index
Harvey's Legal Agent Benchmark (Vals AI)Agentic work 19.6% Not in index
Excel Modeling Benchmark (Vals AI)Agentic work Not measured Not in index
BioMysteryBench (Vals AI)Agentic work 80.0% Not in index
Terminal-Bench-Science (Vals AI)Agentic work 62.9%(Mini-SWE-agent) Not in index
Vals Science IndexAgentic work 62.6% Not in index
Time Horizon Index (Vals AI)Agentic work Not measured Not in index
CUA-bench (Vals AI)Agentic work Not measured Not in index
Robo-bench (Vals AI)Agentic work Not measured Not in index
Public Benefits Bench (Vals AI)General Not measured Not in index
MysteryMechanism (Vals AI)Reasoning 45.0% Not in index
MMMU Pro (Vals AI)Reasoning Not measured Not in index
MATH 500 (Vals AI)Math Not measured Not in index
MGSM (Vals AI)Math Not measured Not in index
SAGE (Vals AI)General Not measured Not in index
LegalBench (Vals AI)Knowledge Not measured Not in index
CaseLaw (Vals AI)Knowledge Not measured Not in index
CorpFin (Vals AI)Knowledge Not measured Not in index
TaxEval (Vals AI)Knowledge Not measured Not in index
MortgageTax (Vals AI)Knowledge Not measured Not in index
MedQA (Vals AI)Knowledge Not measured Not in index
MedCode (Vals AI)Knowledge Not measured Not in index
MedScribe (Vals AI)General Not measured Not in index
Vals Multimodal IndexGeneral Not measured Not in index
AA Math IndexMath Not measured Not in index
GPQA Diamond (Artificial Analysis run)Reasoning Not measured Not in index
Humanity's Last Exam (Artificial Analysis run)Reasoning Not measured Not in index
MMLU Pro (Artificial Analysis run)Knowledge Not measured Not in index
LiveCodeBench (Artificial Analysis run)Coding Not measured Not in index
SciCodeCoding Not measured Not in index
Terminal-Bench Hard (Artificial Analysis run)Coding Not measured Not in index
IFBenchGeneral Not measured Not in index
Long context reasoning (AA-LCR)General Not measured Not in index
τ²-Bench TelecomAgentic work Not measured Not in index
AIME 2025 (Artificial Analysis run)Math Not measured Not in index
Terminal-Bench 4.0 (Artificial Analysis run)Coding 63.6%(xHigh) Not in index
FrontierCode 2.0Coding 44.4%(xHigh) Not in index
HLE-Diamond (Artificial Analysis run)Reasoning 71.3%(xHigh) Not in index
GDPval-AAGeneral Not measured Not in index
AA-OmniscienceKnowledge 54.8(xHigh) Not in index
CritPt v2, hard subset (Artificial Analysis run)Reasoning 61.5%(Medium) Not in index
Terminal-BenchCoding Not measured Not in index
DeepSWECoding Not measured Not in index
FrontierCodeCoding Not measured 15%
FrontierSWECoding Not measured Not in index
GSO-BenchCoding Not measured Not in index
WeirdMLCoding Not measured Not in index
MirrorCodeCoding Not measured Not in index
APEX-AgentsAgentic work 82.9%(xHigh) Not in index
METR time horizonsAgentic work Not measured Not in index
Remote Labor IndexAgentic work Not measured Not in index
GDPvalAgentic work Not measured Not in index
OSWorld 2.0Agentic work Not measured Not in index
BALROGAgentic work Not measured Not in index
DeepResearch BenchAgentic work Not measured Not in index
PostTrainBenchAgentic work Not measured Not in index
GPQA Diamond (Epoch AI run)Reasoning 96.3%(xHigh) Not in index
ARC-AGIReasoning Not measured Not in index
ARC-AGI-2Reasoning Not measured Not in index
Humanity's Last ExamReasoning Not measured Not in index
SimpleBenchReasoning 94.4% Not in index
Chess puzzlesReasoning 83.0%(xHigh) Not in index
Mystery game puzzlesReasoning 85.0%(xHigh) Not in index
EBR-benchReasoning 98.1%(xHigh) Not in index
OTIS Mock AIMEMath 100.0%(xHigh) Not in index
FrontierMath, tiers 1 to 3Math 91.2%(xHigh) Not in index
FrontierMath, tier 4Math 97.6%(xHigh) Not in index
ProofBenchMath Not measured Not in index
MATH level 5Math Not measured Not in index
GSM8KMath Not measured Not in index
SimpleQA VerifiedKnowledge 75.2%(xHigh) Not in index
MMLUKnowledge Not measured Not in index
GeoBenchKnowledge Not measured Not in index
DTBenchGeneral Not measured Not in index
LMCAGeneral Not measured Not in index
Fiction.LiveBenchGeneral Not measured Not in index
Creative writing (Lech Mazur)General Not measured Not in index
CL-benchGeneral Not measured Not in index
Surface Evolver BenchGeneral Not measured Not in index
Furniture assemblyGeneral 89.2%(xHigh) Not in index
ExploitBenchGeneral Not measured Not in index
CybenchCoding Not measured Not in index
TheAgentCompanyAgentic work Not measured Not in index
OSWorldAgentic work Not measured Not in index
SpireClimbAgentic work Not measured Not in index
VPCTReasoning Not measured Not in index
BIG-Bench HardReasoning Not measured Not in index
AI2 Reasoning ChallengeKnowledge Not measured Not in index
HellaSwagGeneral Not measured Not in index
WinoGrandeGeneral Not measured Not in index
PIQAGeneral Not measured Not in index
CL-bench LifeGeneral Not measured Not in index
ALE-BenchCoding Not measured Not in index
Vending-Bench 2Agentic work $16,843 Not in index
Epoch Capabilities IndexIntelligence Not measured Not in index
LMArena TextGeneral 1524(High) Not in index
LMArena Text: codingCoding 1555(High) Not in index
LMArena Text: hard promptsReasoning 1547(High) Not in index
LMArena Text: mathMath 1548(High) Not in index
LMArena Text: instruction followingGeneral 1535(High) Not in index
LMArena Text: creative writingGeneral 1519(High) Not in index
LMArena VisionGeneral Not measured Not in index
LMArena Text: EnglishGeneral 1511(High) Not in index
LMArena Text: non-EnglishGeneral 1524(High) Not in index
LMArena Text: ChineseGeneral 1598(High) Not in index
LMArena Text: GermanGeneral Not measured Not in index
LMArena Text: JapaneseGeneral Not measured Not in index
LMArena Text: KoreanGeneral Not measured Not in index
LMArena Text: FrenchGeneral 1563(High) Not in index
LMArena Text: SpanishGeneral 1505(High) Not in index
LMArena Text: RussianGeneral 1554(High) Not in index
LMArena Text: PolishGeneral Not measured Not in index
LMArena Text: multi-turnGeneral 1525(High) Not in index
LMArena Text: longer queriesGeneral 1533(High) Not in index
LMArena Text: expertReasoning 1560(High) Not in index
LMArena Text: hard prompts in EnglishReasoning 1524(High) Not in index
LMArena Text: software and ITCoding 1542(High) Not in index
LMArena Text: mathematical workMath 1566(High) Not in index
LMArena Text: scienceKnowledge 1549(High) Not in index
LMArena Text: medicine and healthcareKnowledge 1534(High) Not in index
LMArena Text: legal and governmentKnowledge 1518(High) Not in index
LMArena Text: business and financeKnowledge 1493(High) Not in index
LMArena Text: writing and languageGeneral 1517(High) Not in index
LMArena Text: entertainment and mediaGeneral 1495(High) Not in index
LMArena WebDev: ReactCoding 1787(High) Not in index
LMArena WebDev: HTMLCoding 1803(High) Not in index
LMArena WebDev: from an imageCoding Not measured Not in index
LMArena Vision: EnglishGeneral Not measured Not in index
LMArena Vision: ChineseGeneral Not measured Not in index
LMArena Vision: reading textGeneral Not measured Not in index
LMArena Vision: diagramsReasoning Not measured Not in index
LMArena Vision: homeworkReasoning Not measured Not in index
LMArena Vision: humorGeneral Not measured Not in index
LMArena Vision: captioningGeneral Not measured Not in index
LMArena Vision: recognizing thingsGeneral Not measured Not in index
LMArena Vision: creative writingGeneral Not measured Not in index
LMArena SearchAgentic work Not measured Not in index
LMArena DocumentGeneral Not measured Not in index
LMArena Text to ImageImage generation Not measured Not in index
LMArena Image EditImage generation Not measured Not in index
Berkeley Function Calling Leaderboard (BFCL)Agentic work Not measured Not in index
Kagi LLM BenchmarkReasoning Not measured Not in index
BrowseComp-PlusAgentic work Not measured Not in index
SciCodeCoding Not measured Not in index
FACTS Benchmark SuiteKnowledge Not measured Not in index
MathArenaMath Not measured Not in index
MultiChallengeGeneral Not measured Not in index
MCP AtlasAgentic work Not measured Not in index
Humanity's Last Exam (Scale run)Reasoning Not measured Not in index
PRBench FinanceKnowledge Not measured Not in index
PRBench LegalKnowledge Not measured Not in index
SWE-Bench Pro V2Coding Not measured Not in index
SWE Atlas: Codebase QnACoding Not measured Not in index
SWE Atlas: RefactoringCoding Not measured Not in index
SWE Atlas: Test WritingCoding Not measured Not in index
EQ-Bench 4General Not measured Not in index
GameBench 2Coding 98.0(xHigh) Not in index
GoBenchReasoning Not measured Not in index
Terminal-Bench 2.0Coding Not measured Not in index
Terminal-Bench 2.1Coding Not measured Not in index
Terminal-Bench 3.0Coding Not measured Not in index
Terminal-Bench 4.0Coding Not measured Not in index
CursorBench 4.0Coding 65.4%(xhigh) Not in index
Toolathlon-VerifiedAgentic work Not measured Not in index
ProgramBenchCoding Not measured Not in index
τ³-BankingAgentic work Not measured Not in index

Scores by reasoning effort

Some boards test this model at more than one reasoning effort. Rankings use its best run; every run the source published is listed here.

Scores by reasoning effort
BenchmarkEffortScoreSource
AA-Omniscience xHigh · used in rankings 54.8 Artificial Analysis
AA-Omniscience High 53.8 Artificial Analysis
AA-Omniscience Medium 51.1 Artificial Analysis
APEX-Agents xHigh · used in rankings 82.9% Epoch AI benchmarking hub
APEX-Agents High 82.7% Epoch AI benchmarking hub
CritPt v2, hard subset (Artificial Analysis run) Medium · used in rankings 61.5% Artificial Analysis
CritPt v2, hard subset (Artificial Analysis run) xHigh 60.5% Artificial Analysis
CritPt v2, hard subset (Artificial Analysis run) High 59.0% Artificial Analysis
CursorBench 4.0 xhigh · used in rankings 65.4% CursorBench (Cursor)
CursorBench 4.0 high 63.7% CursorBench (Cursor)
CursorBench 4.0 medium 57.7% CursorBench (Cursor)
CursorBench 4.0 low 45.5% CursorBench (Cursor)
FrontierCode 2.0 xHigh · used in rankings 44.4% Artificial Analysis
FrontierCode 2.0 High 43.3% Artificial Analysis
FrontierCode 2.0 Medium 36.9% Artificial Analysis
GameBench 2 xHigh · used in rankings 98.0 DuelLab GameBench 2
GameBench 2 Medium 92.6 DuelLab GameBench 2
GameBench 2 Low 83.3 DuelLab GameBench 2
HLE-Diamond (Artificial Analysis run) xHigh · used in rankings 71.3% Artificial Analysis
HLE-Diamond (Artificial Analysis run) High 68.4% Artificial Analysis
HLE-Diamond (Artificial Analysis run) Medium 63.6% Artificial Analysis
Terminal-Bench 4.0 (Artificial Analysis run) xHigh · used in rankings 63.6% Artificial Analysis
Terminal-Bench 4.0 (Artificial Analysis run) High 61.6% Artificial Analysis
Terminal-Bench 4.0 (Artificial Analysis run) Medium 54.5% Artificial Analysis

What it costs

A list price says what a token costs. The cost per solved task says what a working answer costs, which is the number that decides.

Output price against the market

Gemini 4 Neon pricing
$0.045Dashed line: median $2.00, log scale$180.00
$15.00 per million output tokens is cheaper than 8% of the 341 paid models in the catalog.

Cost per solved task

How it is computed

Neither board that publishes what a SWE-bench Verified attempt costs, Vals AI and swebench.com, has run this model. We leave the number out rather than estimate it.

LiveBench, on its own coding tasks, publishes what one task cost this model: $6.919, at a score of 89.7%. LiveBench counts that as $7.714 per successful task. Different tasks from SWE-bench, so it is not our number and does not enter its ranking. LiveBench coding

Nearly as good as Gemini 4 Neon, for less

What the boards say about its runs

Beside a result, many boards publish what the run cost, how long it took, the tokens it spent or the votes behind a rating. Each figure as its board printed it, one board per row.

What each board publishes about the runs behind Google: Gemini 4 Neon's results
BoardResultWhat the run costAlso published
LiveBench coding 89.7% $6.9191 per task Output tokens per task 120,302
Vals Index 70.9% $7.40 per test Nothing else
LiveBench reasoning 89.6% $0.3379 per task Output tokens per task 21,801
LiveBench data analysis 85.3% $1.0783 per task Output tokens per task 62,808
LiveBench research 94.5% $6.8955 per task Output tokens per task 109,487
LiveBench browser 95.6% $9.9409 per task Output tokens per task 26,101
LiveBench long agentic 82.6% $13.1205 per task Output tokens per task 311,909
LiveBench design 81.5% $1.9565 per task Output tokens per task 81,681
Terminal-Bench 4 (Vals AI) 61.6%(Mini-SWE-agent) $6.42 per test Nothing else
Vibe Code Bench (Vals AI) 93.7%(OpenHands) $6.76 per test Nothing else
Tax Agent Bench (Vals AI) 54.9% $4.95 per test Nothing else
Legal Research Bench (Vals AI) 56.3% $7.37 per test Nothing else
38 more boards
What each board publishes about the runs behind Google: Gemini 4 Neon's results, continued
Harvey's Legal Agent Benchmark (Vals AI) 19.6% $5.97 per test Nothing else
BioMysteryBench (Vals AI) 80.0% $4.41 per test Nothing else
Terminal-Bench-Science (Vals AI) 62.9%(Mini-SWE-agent) $13.74 per test Nothing else
Vals Science Index 62.6% $8.26 per test Nothing else
MysteryMechanism (Vals AI) 45.0% $6.63 per test Nothing else
GameBench 2 98.0(xHigh) $1.8576 per program First-try builds 87.5% · Failed programs 0.0% · Matches 5,358 · Wins 5,171 · Draws 51 · Losses 136
LMArena WebDev 1765 Elo(High) Not published Votes 2,305 · Interval (±) ±13.7
Design Arena: UI components 1349 Elo Not published Battles 131
Design Arena: game dev 1394 Elo Not published Battles 641
Design Arena: 3D 1490 Elo Not published Battles 580
Design Arena: all code categories 1391 Elo Not published Battles 1,714
APEX-Agents 82.9%(xHigh) Not published Mean score 88.9%
LMArena Text 1524 Elo(High) Not published Votes 5,555 · Interval (±) ±8.5
LMArena Text: coding 1555 Elo(High) Not published Votes 1,288 · Interval (±) ±17.2
LMArena Text: hard prompts 1547 Elo(High) Not published Votes 3,508 · Interval (±) ±10.6
LMArena Text: math 1548 Elo(High) Not published Votes 212 · Interval (±) ±40.1
LMArena Text: instruction following 1535 Elo(High) Not published Votes 1,918 · Interval (±) ±14.1
LMArena Text: creative writing 1519 Elo(High) Not published Votes 1,204 · Interval (±) ±18.6
LMArena Text: English 1511 Elo(High) Not published Votes 1,843 · Interval (±) ±14.1
LMArena Text: non-English 1524 Elo(High) Not published Votes 3,701 · Interval (±) ±10.3
LMArena Text: Chinese 1598 Elo(High) Not published Votes 426 · Interval (±) ±30.2
LMArena Text: French 1563 Elo(High) Not published Votes 183 · Interval (±) ±45.5
LMArena Text: Spanish 1505 Elo(High) Not published Votes 179 · Interval (±) ±44.2
LMArena Text: Russian 1554 Elo(High) Not published Votes 970 · Interval (±) ±19.6
LMArena Text: multi-turn 1525 Elo(High) Not published Votes 815 · Interval (±) ±22.1
LMArena Text: longer queries 1533 Elo(High) Not published Votes 2,348 · Interval (±) ±13.1
LMArena Text: expert 1560 Elo(High) Not published Votes 423 · Interval (±) ±29.3
LMArena Text: hard prompts in English 1524 Elo(High) Not published Votes 1,217 · Interval (±) ±17.3
LMArena Text: software and IT 1542 Elo(High) Not published Votes 1,944 · Interval (±) ±13.8
LMArena Text: mathematical work 1566 Elo(High) Not published Votes 247 · Interval (±) ±37.7
LMArena Text: science 1549 Elo(High) Not published Votes 827 · Interval (±) ±21.2
LMArena Text: medicine and healthcare 1534 Elo(High) Not published Votes 323 · Interval (±) ±34.0
LMArena Text: legal and government 1518 Elo(High) Not published Votes 424 · Interval (±) ±29.2
LMArena Text: business and finance 1493 Elo(High) Not published Votes 891 · Interval (±) ±20.1
LMArena Text: writing and language 1517 Elo(High) Not published Votes 1,635 · Interval (±) ±15.6
LMArena Text: entertainment and media 1495 Elo(High) Not published Votes 1,423 · Interval (±) ±16.7
LMArena WebDev: React 1787 Elo(High) Not published Votes 1,444 · Interval (±) ±17.5
LMArena WebDev: HTML 1803 Elo(High) Not published Votes 400 · Interval (±) ±35.2

Google Gemini 4 Neon API providers

On these lists

Compare with models of a similar score

All 110 comparisons for Gemini 4 Neon