Skip to main content

OpenAIGoogle DeepMindLLM matchup

GPT-6.2 Astravs.Gemini 4 Neon

Which one is better?

GPT-6.2 Astra and Gemini 4 Neon perform about equally well, but GPT-6.2 Astra costs 2.7 times as much.

level
IntelligenceOverall index, 100 points = average
126.6

GPT-6.2 Astra

Rank 4

127.1

Gemini 4 Neon

Rank 3

Gemini 4 Neon cheaper
Price$ per 1M tokens
$16

GPT-6.2 Astra

$6

Gemini 4 Neon

With more than 272,000 input tokens, GPT-6.2 Astra costs more: $16 input and $60 output per 1M tokens.

Choose GPT-6.2 Astra for

Coding (not certain)General (not certain)

Choose Gemini 4 Neon for

FinanceLegal

Where is the difference?

Strengths at a glance

GPT-6.2 AstraGemini 4 Neon

The median of all ranked models is below 105 in every area, so it sits in the centre of the net.

115125135105CodingslightlyGPT-6.2 AstraReasoninglevelAgentslevelGeneralslightlyGPT-6.2 AstraFinanceGemini 4 NeonLegalGemini 4 Neon
CodingGPT-6.2 Astra slightly ahead
GPT-6.2 Astra131.1Rank 1
Gemini 4 Neon126.2Rank 5
Median of ranked models84.3

95 % range: GPT-6.2 Astra 121.7 to 140.5, Gemini 4 Neon 120.1 to 132.2.

AreaGPT-6.2 AstraGemini 4 Neon

The same values and verdicts as the bars below. In the close-up the net starts at 105 points, just below the weaker model's lowest value; the reference group average is 100. Not in the net because a model has no own value there or the data are too thin: Mathematics, Knowledge, Multimodal, Health.

Gemini 4 Neon is ahead in 2 of 9 areas, GPT-6.2 Astra in none.

GPT-6.2 Astra slightly ahead: 2level: 2too little data: 3Gemini 4 Neon ahead: 2

Areas in the overall index
Codingleaning GPT-6.2 Astranot certain · 7 tests each
Generalleaning GPT-6.2 Astranot certain · 3 tests each
Reasoninglevel124.2 to 121.3
Agentslevel122.2 to 122.6
Knowledgetoo little data133.1 to 135.6
Mathematicsonly GPT-6.2 Astra has a score
Multimodalonly GPT-6.2 Astra has a score
Specialist fields
FinanceGemini 4 Neon +16111.3 to 127.7
LegalGemini 4 Neon +8118.8 to 126.6

Bar length shows the gap in index points (full width 15 points, one standard deviation of the reference group). Solid means clearly ahead, hatched not certain yet, a thin grey bar level.

GPT-6.2 Astra vs. Gemini 4 Neon at the same effort

Both models measured at the same reasoning efforts. This shows whether a lead comes from the model or only from more compute.

lowtied
1.7 points gap across 9 shared test series
mediumtied
0.6 points gap across 9 shared test series
hightied
0.7 points gap across 9 shared test series
xhightied
0.5 points gap across 9 shared test series
Overall index (max vs. xhigh)level
GPT-6.2 Astra 126.6 vs. Gemini 4 Neon 127.1

The gap compares only the Artificial Analysis test series in which both models are measured at that effort. Tied below 3 points, ahead only beyond two standard errors.

Which model for which task?

Everyday tasks, translated from the area values above. Each row names the values it rests on.

Writing and debugging codeleaning GPT-6.2 AstraCoding 131.1 to 126.2
Maths and data analysislevelMathematics without a score for Gemini 4 Neon, Reasoning 124.2 to 121.3
Facts and researchopenKnowledge 133.1 to 135.6 (too little data)
Agents and workflowslevelAgents 122.2 to 122.6
Many requests on a small budgetGemini 4 Neon$16 to $6 per 1M tokens
Finance, legal and medical textsGemini 4 NeonFinance 111.3 to 127.7, Legal 118.8 to 126.6
Images, screenshots and PDFsopenMultimodal without a score for Gemini 4 Neon

GPT-6.2 Astra vs. Gemini 4 Neon in our own test

The same prompt for both models. We show the latest attempt without fixes; it does not count toward the overall index.

Pelican on a bicycle

Prompt

Generate an SVG of a pelican riding a bicycle

Task origin: Simon Willison

GPT-6.2 Astra

Not tested yet

Gemini 4 Neon

Not tested yet

GPT-6.2 Astra and Gemini 4 Neon in the overall index

Every model with a place in the overall index, both models of this matchup highlighted. You can add or hide other models.

Overall index (Gradually Meta-Score v6)

20 of 99 models

Claude Fable 5.2max130.3
Gemini 4 Carbonhigh127.8
Gemini 4 Neonxhigh127.1
GPT-6.2 Astramax126.6
Claude Opus 5.5max125.0
Claude Fable 5.1max124.1
Claude Sonnet 5.5max121.9
Gemini 4 Argonhigh121.5
GPT-6 Astramax119.6
GPT-6.2 Solmax119.2
Muse Wonder Fullmax117.9
GPT-6.1 Solmax117.3
Claude Opus 5max117.1
Muse Wondermax113.5
Claude Fable 5max112.8
Kimi K3.1 Promax112.4
Muse Ember 1.0max111.1
GPT-5.6 Solmax109.5
GPT-6 Solmax108.3
Gemini 3.7 Flashhigh107.8

All data side by side

The better value is bold. For the overall index only when the gap is clear.

MetricGPT-6.2 AstraGemini 4 Neon
Intelligence
Overall indexpoints, 100 = reference group average, higher is better126.6127.1
RankRank 4 of 99Rank 3 of 99
Price per 1M tokens
Inputwhat you send$8above 272,000 input tokens: $16$3
Outputwhat the model writes$40above 272,000 input tokens: $60$15
Speed
Tokens per secondn/an/a
Technical
Context windowsource-reported context limit1.05 M tokens≈ 1,580 A4 pages1.05 M tokens≈ 1,570 A4 pages
Knowledge cutoffMay 18, 2026March 2026
ReleasedNov 17, 2026Nov 18, 2026
AvailabilityActiveActive
Open weightsnono
  • GPT-6.2 Astra, 12-hour cache write price: The table lists the 30-minute rate. For 12-hour retention, OpenAI charges $20 per 1M tokens at the Standard rate through 272,000 input tokens. Source
  • GPT-6.2 Astra, Long-context pricing: Above 272,000 input tokens, the higher rates apply to the entire request. Source
  • GPT-6.2 Astra, Batch, Fast, and Ultrafast: Batch costs half the Standard rate and Fast costs twice the Standard rate. Ultrafast costs three times the Standard rate through February 28, 2027 and six times the rate from March 1, 2027. The change follows Pacific time. Source
  • Gemini 4 Neon, Billing: The output price includes reasoning tokens. Cache hits are billed through implicit caching. Source
  • Gemini 4 Neon, Batch, Flex, and Priority: Batch and Flex cost half the Standard rate. Priority costs 1.8 times the Standard rate. These multipliers apply to input, output, and cache hits. Source

Prices without caching or batch discounts · Page count as a rule of thumb from Artificial Analysis

Is there a better choice?

Cheaper than Gemini 4 NeonGemini 4 Carbon

Overall index 127.8, rank 2$2 input and $10 output per 1M tokens

Gemini 4 Neon vs. Gemini 4 Carbon
Best model in the overall indexClaude Fable 5.2

Overall index 130.3, rank 1$8 input and $40 output per 1M tokens

Gemini 4 Neon vs. Claude Fable 5.2

Shared benchmarks

Single results from the same benchmark version, task and metric. A value is bold only with the same test setup and a clear gap.

A: GPT-6.2 AstraB: Gemini 4 Neon

AA-BriefcaseOverall · higher is betterGPT-6.2 Astra: 1,594.4 eloGemini 4 Neon: 1,699.2 eloArtificial AnalysisAs of 11/20/2026
AA-OmniscienceOverall · higher is betterGPT-6.2 Astra: 48.3 pointsGemini 4 Neon: 54.8 pointsArtificial AnalysisAs of 11/20/2026
AA-Omniscience AccuracyOverall · higher is betterGPT-6.2 Astra: 67.2%Gemini 4 Neon: 60.5%Artificial AnalysisAs of 11/20/2026
AA-Omniscience Hallucination RateOverall · lower is betterGPT-6.2 Astra: 57.7%Gemini 4 Neon: 14.4%Artificial AnalysisAs of 11/20/2026
AutomationBench-AAOverall · higher is betterGPT-6.2 Astra: 76.1%Gemini 4 Neon: 79.3%Artificial AnalysisAs of 11/20/2026
BioMysteryBenchOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 80.7%Gemini 4 Neon: 80.0%Vals AIAs of 11/18/2026
CritPt v2 (Hard Subset)Overall · higher is betterGPT-6.2 Astra: 70.5%Gemini 4 Neon: 60.5%Artificial AnalysisAs of 11/20/2026
EMB v1.1Overall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 76.0%Gemini 4 Neon: 76.5%Vals AIAs of 11/18/2026
Finance Agent v2.1Overall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 60.7%Gemini 4 Neon: 68.6%Vals AIAs of 11/18/2026
FrontierCodeOverall · higher is betterGPT-6.2 Astra: 49.0%Gemini 4 Neon: 44.4%Artificial AnalysisAs of 11/20/2026
Show 54 more benchmarks
GDP.pdfOverall · higher is betterGPT-6.2 Astra: 28.0%Gemini 4 Neon: 28.0%Artificial AnalysisAs of 11/20/2026
HLAB-Vals-1.1Overall · Task fully resolved · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 16.3%Gemini 4 Neon: 19.6%Vals AIAs of 11/18/2026
HLE-DiamondOverall · higher is betterGPT-6.2 Astra: 67.3%Gemini 4 Neon: 71.3%Artificial AnalysisAs of 11/20/2026
Legal Research BenchOverall · All-pass · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 42.8%Gemini 4 Neon: 56.3%Vals AIAs of 11/18/2026
MysteryMechanismOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 53.2%Gemini 4 Neon: 45.0%Vals AIAs of 11/18/2026
Tax Agent BenchOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 21.8%Gemini 4 Neon: 54.9%Vals AIAs of 11/18/2026
Terminal-BenchOverall · higher is betterGPT-6.2 Astra: 64.1%Gemini 4 Neon: 63.6%Artificial AnalysisAs of 11/20/2026
Terminal-Bench 4.0Overall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 63.1%Gemini 4 Neon: 61.6%Vals AIAs of 11/18/2026
Terminal-Bench-ScienceOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 71.4%Gemini 4 Neon: 62.9%Vals AIAs of 11/18/2026
Terminal-Bench-ScienceOverall · higher is betterGPT-6.2 Astra: 78.6%Gemini 4 Neon: 70.0%Artificial AnalysisAs of 11/20/2026
Vals Finance IndexOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 57.3%Gemini 4 Neon: 68.3%Vals AIAs of 11/18/2026
Vals IndexOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 67.7%Gemini 4 Neon: 70.9%Vals AIAs of 11/18/2026
Vals Legal IndexOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 69.7%Gemini 4 Neon: 78.4%Vals AIAs of 11/18/2026
Vals Science IndexOverall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 68.4%Gemini 4 Neon: 62.6%Vals AIAs of 11/18/2026
Vibe Code Bench v1.1Overall · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 89.9%Gemini 4 Neon: 93.7%Vals AIAs of 11/18/2026
BioMysteryBenchHard · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 41.2%Gemini 4 Neon: 37.3%Vals AIAs of 11/18/2026
EMB v1.1Comps · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 83.1%Gemini 4 Neon: 82.6%Vals AIAs of 11/18/2026
Finance Agent v2.1Adjustments · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 55.7%Gemini 4 Neon: 72.8%Vals AIAs of 11/18/2026
HLAB-Vals-1.1Antitrust Competition · Criteria passed · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 92.6%Gemini 4 Neon: 96.0%Vals AIAs of 11/18/2026
Legal Research BenchAdministrative / Regulatory · All-pass · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 59.1%Gemini 4 Neon: 71.2%Vals AIAs of 11/18/2026
Tax Agent BenchControversy & Precedence Analysis · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 8.7%Gemini 4 Neon: 40.0%Vals AIAs of 11/18/2026
Terminal-Bench 4.0Hardware · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 80.0%Gemini 4 Neon: 73.3%Vals AIAs of 11/18/2026
Terminal-Bench-ScienceEarth Sciences · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 87.5%Gemini 4 Neon: 87.5%Vals AIAs of 11/18/2026
Vals Finance IndexEMB v1.1 · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 76.0%Gemini 4 Neon: 76.5%Vals AIAs of 11/18/2026
Vals IndexCode Migration v1.1 · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 83.6%Gemini 4 Neon: 75.8%Vals AIAs of 11/18/2026
Vals Legal IndexBusiness & Commercial · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 64.7%Gemini 4 Neon: 75.6%Vals AIAs of 11/18/2026
Vals Science IndexBioMysteryBench · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 80.7%Gemini 4 Neon: 80.0%Vals AIAs of 11/18/2026
BioMysteryBenchHuman-solvable · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 90%Gemini 4 Neon: 90%Vals AIAs of 11/18/2026
EMB v1.1Dataroom · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 88.6%Gemini 4 Neon: 88.2%Vals AIAs of 11/18/2026
Finance Agent v2.1All-Pass · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 47.2%Gemini 4 Neon: 58.2%Vals AIAs of 11/18/2026
HLAB-Vals-1.1Antitrust Competition · Task fully resolved · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 10%Gemini 4 Neon: 30%Vals AIAs of 11/18/2026
Legal Research BenchAdministrative / Regulatory · Weighted partial credit (weighted pass) · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 87.9%Gemini 4 Neon: 94.3%Vals AIAs of 11/18/2026
Tax Agent BenchControversy & Precedence Analysis (weighted pass) · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 55.9%Gemini 4 Neon: 70.7%Vals AIAs of 11/18/2026
Terminal-Bench 4.0Media · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 75%Gemini 4 Neon: 75%Vals AIAs of 11/18/2026
Terminal-Bench-ScienceEngineering Sciences · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 55.6%Gemini 4 Neon: 33.3%Vals AIAs of 11/18/2026
Vals Finance IndexFinance Agent v2.1 · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 60.7%Gemini 4 Neon: 68.6%Vals AIAs of 11/18/2026
Vals IndexEMB v1.1 · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 76.0%Gemini 4 Neon: 76.5%Vals AIAs of 11/18/2026
Vals Legal IndexCivil Negligence · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 62.1%Gemini 4 Neon: 72.6%Vals AIAs of 11/18/2026
Vals Science IndexMysteryMechanism · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 53.2%Gemini 4 Neon: 45.0%Vals AIAs of 11/18/2026
EMB v1.1DCF · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 66.1%Gemini 4 Neon: 68.8%Vals AIAs of 11/18/2026
Finance Agent v2.1Comparables · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 48.8%Gemini 4 Neon: 60.1%Vals AIAs of 11/18/2026
HLAB-Vals-1.1Arbitration International Dispute Resolution · Criteria passed · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 94.6%Gemini 4 Neon: 97.1%Vals AIAs of 11/18/2026
Legal Research BenchBusiness & Commercial · All-pass · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 34.9%Gemini 4 Neon: 55.8%Vals AIAs of 11/18/2026
Tax Agent BenchCurrent & Temporal Analysis · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 34.1%Gemini 4 Neon: 56.5%Vals AIAs of 11/18/2026
Terminal-Bench 4.0ML · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 57.6%Gemini 4 Neon: 81.8%Vals AIAs of 11/18/2026
Terminal-Bench-ScienceLife Sciences · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 68.4%Gemini 4 Neon: 63.2%Vals AIAs of 11/18/2026
Vals Finance IndexTax Agent Bench · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 21.8%Gemini 4 Neon: 54.9%Vals AIAs of 11/18/2026
Vals IndexFinance Agent v2.1 · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 60.7%Gemini 4 Neon: 68.6%Vals AIAs of 11/18/2026
Vals Legal IndexCriminal · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 68.4%Gemini 4 Neon: 74.5%Vals AIAs of 11/18/2026
Vals Science IndexTerminal-Bench-Science · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 71.4%Gemini 4 Neon: 62.9%Vals AIAs of 11/18/2026
EMB v1.1LBO · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 65.3%Gemini 4 Neon: 66.3%Vals AIAs of 11/18/2026
Finance Agent v2.1Disclosure Analysis · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 63.5%Gemini 4 Neon: 85.2%Vals AIAs of 11/18/2026
HLAB-Vals-1.1Arbitration International Dispute Resolution · Task fully resolved · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 10%Gemini 4 Neon: 30%Vals AIAs of 11/18/2026
Legal Research BenchBusiness & Commercial · Weighted partial credit (weighted pass) · higher is betterTest setup not documented as identicalGPT-6.2 Astra: 80.5%Gemini 4 Neon: 91.9%Vals AIAs of 11/18/2026

More comparisons

Compare GPT-6.2 Astra and Gemini 4 Neon with the models people compare them with most often.

Compare another matchup