Skip to main content

Google DeepMindLanguage modelActive

Gemini 4 Argon

Gemini 4 Argon is in the upper group: Rank 6 of 95 in the overall index. Through the API it is fairly expensive.

IntelligenceRank 6 of 95 · upper group

121.4Overall indexhigher is better

SpeedRank 92 of 99

36Tokens per secondhigher is better

PriceRank 76 of 102

Input $2 · Output $10

$4per 1M tokenslower is better

ContextRank 12 of 111

1.05Mtokens in context≈ 1,570 A4 pages

Popularityno data yet

n/ano search or usage data

Compare with another model

Summary

It is fairly expensive: $4 per 1M tokens (input and output blended); a typical model costs $1.50. At 36 tokens per second it writes slower than a typical model (70).

Google announced Gemini 4 Argon on September 30, 2026 and made it generally available through the Gemini API and Google AI Studio on October 13. Its context window holds 1,048,576 tokens shared across input, thinking, and output.

Synchronous output is capped at 65,536 tokens. Long Decode can produce up to 1,048,576 output tokens minus the input. Google also highlights finding, validating, and patching security vulnerabilities as a focus.

Prices

Choose a price view

1 offer, prices in US dollars per 1M tokens, blended from 3 parts input and 1 part output.

ProviderInputOutputCache readBlended
Google DeepMinddirect$2$10$0.10$4

Direct price according to the developer, checked Oct 25, 2026. Open the LLM price index

More price tiers and rules
Standard pricing through November 30, 2026
Per 1 million tokens: $2.00 input, $10.00 output including thinking, and $0.10 cached input.Source
Standard pricing from December 1, 2026
Per 1 million tokens: $4.00 input, $20.00 output including thinking, and $0.20 cached input.Source
Caching
Implicit caching only; explicit caching is not supported.Source

At Google DeepMind from $19.99 a month (Google AI Pro, with limited credits). Through other providers from $6 a month (ElevenLabs Starter, with limited credits). 30 plans in Germany.

Filter plans
PlanMonthly list priceNote
Google AI Pro$19.99developer's own plan, with limited credit
Google AI Ultra 5x$99.99developer's own plan, with limited credit
Google AI Ultra 20x$199.99developer's own plan, with limited credit
ElevenLabs Freefreewith limited credit
Lovable Freefreewith limited credit
Zed Studentfree
ElevenLabs Starter$6with limited credit
Zed Pro$10
GitHub Copilot Business$19with limited credit, for teams, per seat
Command Code Pro$20with limited credit
ElevenLabs Creator$22with limited credit
Lovable Pro$25with limited credit
Langdock Core Standard€29for teams, per seat
GitHub Copilot Enterprise$39with limited credit, for enterprises, per seat
GitHub Copilot Pro+$39with limited credit
Genspark for Business Team$40with limited credit, for teams, per seat
Langdock Work Standard€39for teams, per seat
Lovable Business$50with limited credit, for teams
ElevenLabs Pro$99with limited credit
Command Code Max 10×$100with limited credit
GitHub Copilot Max$100with limited credit
Langdock Core Max€119for teams, per seat
Langdock Work Max€129for teams, per seat
Command Code Max 20×$200with limited credit
Perplexity Max$200with limited credit
ElevenLabs Scale$299with limited credit, for teams
ElevenLabs Business$990with limited credit, for teams
ElevenLabs Enterprisen/afor enterprises
Genspark for Business Enterprisen/awith limited credit, for enterprises, per seat
Lovable Enterprisen/afor enterprises

All plans in the AI subscription comparison

Gemini 4 Argon benchmarks

Overall index (Gradually Meta-Score v6)

20 of 95 models

Speed

20 of 99 models

GPT OSS 120B165tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Claude Haiku 5.5160tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Nemotron 3.5 Lightning147tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Gemini 3.5 Flash-Lite147tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Gemini 3.6 Flash141tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
DeepSeek V4.1 Flash 1017140tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
DeepSeek-V4.1-Flash136tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Gemini 3.1 Flash-Lite132tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
MiniMax M2.7132tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Inkling Small131tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Gemini 3.5 Flash122tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Ling 3.0 Flash119tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Inkling115tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Mistral Medium 3.5109tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Claude Sonnet 5.5107tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Mistral Small 4107tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Gemini 3.1 Pro Preview107tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
GPT-5.6 Luna105tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Muse Spark 1.3104tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20
Gemini 4 Argon36tokens/s · Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token · Retrieved 2026-11-20

Output tokens per second, measured by OpenRouter

Intelligence vs. price

most attractive area: stronger and cheaper than half of the models shownbest intelligence for the price

100120140$0.20$1$5$20DeepSeek-V4.1-Flash: 100.0, $0.26GLM-5.3: 101.0, $2.15Mistral Large 4: 101.0, $1.05GPT-5.6 Terra: 101.2, $4.50MiMo V2.6 Pro: 101.4, $0.54GPT-5.5: 101.4, $11.25Kimi K3: 101.4, $6Grok 4.6: 102.2, $3Claude Haiku 5.5: 102.6, $0.20Grok 4.7: 103.6, $3Qwen 3.8 Max 0902: 104.0, $2.48Muse Spark 1.3: 104.3, $2Gemini 3.8 Flash: 106.4, $1.50GPT-6.1 Terra: 106.6, $2DeepSeek V4.1 Flash 1017: 106.6, $0.26GPT-6 Sol: 108.3, $4GPT-5.6 Sol: 109.4, $8Muse Ember 1.0: 110.8, $0.20Kimi K3.1 Pro: 112.6, $4Claude Fable 5: 112.8, $20Muse Wonder: 113.5, $8Claude Opus 5: 117.1, $10GPT-6.1 Sol: 117.4, $4GPT-6 Astra: 119.6, $20Claude Sonnet 5.5: 121.8, $4Claude Fable 5.1: 124.1, $20Claude Opus 5.5: 125.1, $8Gemini 4 Carbon: 127.9, $4Claude Fable 5.2: 130.5, $16Gemini 4 Argon: 121.4, $4Gemini 4 ArgonClaude Fable 5.2Gemini 4 CarbonPrice per 1M tokens, blended (log scale)Overall index
More models in the chart (27)
  • Claude Opus 5.5125.1 · $8
  • Claude Fable 5.1124.1 · $20
  • Claude Sonnet 5.5121.8 · $4
  • GPT-6 Astra119.6 · $20
  • GPT-6.1 Sol117.4 · $4
  • Claude Opus 5117.1 · $10
  • Muse Wonder113.5 · $8
  • Claude Fable 5112.8 · $20
  • Kimi K3.1 Pro112.6 · $4
  • Muse Ember 1.0110.8 · $0.20
  • GPT-5.6 Sol109.4 · $8
  • GPT-6 Sol108.3 · $4
  • DeepSeek V4.1 Flash 1017106.6 · $0.26
  • GPT-6.1 Terra106.6 · $2
  • Gemini 3.8 Flash106.4 · $1.50
  • Muse Spark 1.3104.3 · $2
  • Qwen 3.8 Max 0902104.0 · $2.48
  • Grok 4.7103.6 · $3
  • Claude Haiku 5.5102.6 · $0.20
  • Grok 4.6102.2 · $3
  • Kimi K3101.4 · $6
  • GPT-5.5101.4 · $11.25
  • MiMo V2.6 Pro101.4 · $0.54
  • GPT-5.6 Terra101.2 · $4.50
  • Mistral Large 4101.0 · $1.05
  • GLM-5.3101.0 · $2.15
  • DeepSeek-V4.1-Flash100.0 · $0.26

Gemini 4 Argon by task area

Points, 100 = reference group average · higher is better

KnowledgeRank 2 of 32127.1
CodingRank 7 of 118123.5
HealthRank 3 of 72122.8
FinanceRank 3 of 114122.4
GeneralRank 7 of 71120.2
LegalRank 7 of 82120.1
ReasoningRank 6 of 78119.2
MathematicsRank 4 of 48118.4
AgentsRank 7 of 102113.8

No tests yet: Multimodal.

Best LLM by task:CodingReasoningMathKnowledgeAI agentsGeneral tasksFinanceLegalMedicine

Text Arena

How people rate the answers in blind comparisons, Elo points with style control · higher is better

20 of 40 models

as of November 19, 2026 · Source: LMArena

WebDev Arena

How people rate generated websites and web apps in blind comparisons, Elo points · higher is better

20 of 43 models

as of November 19, 2026 · Source: LMArena

Terminal-Bench 4.0

Share of tasks the model solves as an agent in the terminal, in percent · higher is better

20 of 41 models

as of November 19, 2026 · Source: Vals AI

Vibe Code Bench

Share of apps the model builds correctly from a description, in percent · higher is better

20 of 54 models

as of November 19, 2026 · Source: Vals AI

Show all 977 published benchmark results

Page 1 of 82

Published benchmarks

Autoform-Hard

Overall

Gemini 4 Argon48.21%

higher is betterAs of November 19, 2026Scale: 0% - 100%

Source: Vals AI
gradually.ai
Methodology and context
Comparison cohort
vals-autoform-hard:0.2:overall
Metric
accuracy
Scale
0-100
Participants
2
Methodology and context
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-11-18. Reasoning effort: high. Maximum output tokens: 1048576. Latency in seconds: 6487.894. Standard error: 6.738. Cost per task in USD: 25.736314. Source ID: vals-ai. Evidence capture ID: vals-autoform_hard-2026-11-19-cdf885d8f740. Source payload hash: cdf885d8f740170eee59379e6146d9256add098c21843214c75e52ad89b1bb62. Parser version: vals-astro-v4. Imported at: 2026-11-19
Source: Vals AI

Other tasks

Vals Index

Code Migration v1.1

Gemini 4 Argon74.43%

higher is betterAs of November 7, 2026Scale: 0% - 100%

Source: Vals AI
gradually.ai
Methodology and context
Comparison cohort
vals-vals-index:2.2:code_migration
Metric
benchmark score
Scale
0-100
Participants
2
Methodology and context
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-05. Reasoning effort: high. Maximum output tokens: 1048576. Latency in seconds: 8400.232. Standard error: 4.512. Cost per task in USD: 44.644196. Source ID: vals-ai. Evidence capture ID: vals-vals_index-2026-11-07-d2a0c2d8009e. Source payload hash: d2a0c2d8009ee84f7dafadf2065edeaffb4335c59abae4bae8e0f4d59cff89af. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source: Vals AI

Vals Router Index

Sub-taskGemini 4 Argon (This model)Claude Sonnet 5.5GPT-6 AstraGPT-6.1 SolMuse WonderDeepSeek V4.1 Flash 1017Muse Wonder FullGLM-5.3Grok 4.7GLM-5.3-FlashClaude Haiku 5.5Claude Opus 5.5GPT-6 LunaGemini 3.8 FlashKimi K3
Code Migration74.43%87.37% (best value in this row)78.03%75.14%69.78%68.5%61.65%54.08%51.82%50.63%50.2%47.49%45.45%38.17%22.67%
Coding74.64%80.8% (best value in this row)75.74%73.04%67.46%62.68%68.58%57.03%55.59%39.59%59%62.97%46.91%45.51%41.6%
EMB74.83%75.9% (best value in this row)72.68%71.09%72.5%69.17%75%69.11%71.5%66.25%65.48%75.84%69.95%72.87%71.26%
Finance Agent67.66% (best value in this row)61.7%60.43%61.82%65.55%62.93%67.58%65%60.59%60.59%65.37%66.76%56.84%65.76%60.01%
HLAB27.08% (best value in this row)10.42%7.08%6.25%23.75%14.17%23.33%14.58%18.33%10.42%4.17%9.17%4.17%12.08%12.92%
Legal Research54.81% (best value in this row)44.23%39.42%38.46%50.48%45.19%53.37%49.04%47.12%45.19%43.27%50.48%30.29%38.94%46.15%
Full Index69.24% (best value in this row)69.03%65.02%63.94%64.88%61.26%66.79%59.73%58.56%50.86%58.71%64.16%52.32%55.95%52.96%
Tax Agent Bench44.56%40.92%20.72%21.75%36.27%44.35%43%39.7%34.2%31.55%31.92%45.57% (best value in this row)19.95%32.12%33.37%
Terminal-Bench 4.057.58%62.63% (best value in this row)59.6%55.05%49.5%31.82%58.08%38.89%28.79%37.37%36.36%58.08%13.64%19.7%17.17%
Vibe Code Bench91.91%92.39% (best value in this row)89.59%88.93%83.09%87.73%86%78.13%86.18%30.76%90.44%83.34%81.65%78.65%84.97%

higher is betterAs of November 7, 2026Source: Vals AIbars scaled within each row, not across rows

Methodology and context
Metric
accuracy

Recorded measurement details

Autoform-Hard: Overall (Show measurement, test conditions, and source)
Source value
48.214
Metric
accuracy
Unit
%
Benchmark version
0.2
Category
math
Task
overall
Task label
Overall
Direction
Higher is better
Comparison cohort
vals-autoform-hard:0.2:overall
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-19
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-11-18. Reasoning effort: high. Maximum output tokens: 1048576. Latency in seconds: 6487.894. Standard error: 6.738. Cost per task in USD: 25.736314. Source ID: vals-ai. Evidence capture ID: vals-autoform_hard-2026-11-19-cdf885d8f740. Source payload hash: cdf885d8f740170eee59379e6146d9256add098c21843214c75e52ad89b1bb62. Parser version: vals-astro-v4. Imported at: 2026-11-19
Source Vals AI
Vals Index: Code Migration v1.1 (Show measurement, test conditions, and source)
Source value
74.429
Metric
benchmark score
Unit
%
Benchmark version
2.2
Category
aggregate
Task
code_migration
Task label
Code Migration v1.1
Direction
Higher is better
Comparison cohort
vals-vals-index:2.2:code_migration
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Provider reported by source: Google. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-05. Reasoning effort: high. Maximum output tokens: 1048576. Latency in seconds: 8400.232. Standard error: 4.512. Cost per task in USD: 44.644196. Source ID: vals-ai. Evidence capture ID: vals-vals_index-2026-11-07-d2a0c2d8009e. Source payload hash: d2a0c2d8009ee84f7dafadf2065edeaffb4335c59abae4bae8e0f4d59cff89af. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Code Migration (Show measurement, test conditions, and source)
Source value
74.429
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
code_migration
Task label
Code Migration
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:code_migration
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 8400.232. Standard error: 4.512. Cost per task in USD: 44.644196. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Coding (Show measurement, test conditions, and source)
Source value
74.637
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
coding
Task label
Coding
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:coding
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 4620.900000000001. Standard error: 1.8048989199152154. Cost per task in USD: 23.46403533333333. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: EMB (Show measurement, test conditions, and source)
Source value
74.828
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
emb
Task label
EMB
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:emb
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 2305.821. Standard error: 2.494. Cost per task in USD: 11.8029. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Finance Agent (Show measurement, test conditions, and source)
Source value
67.658
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
finance_agent
Task label
Finance Agent
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:finance_agent
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 793.03. Standard error: 2. Cost per task in USD: 4.315759. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: HLAB (Show measurement, test conditions, and source)
Source value
27.083
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
legal_agent_benchmark
Task label
HLAB
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:legal_agent_benchmark
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 1755.397. Standard error: 3.738. Cost per task in USD: 10.54836. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Legal Research (Show measurement, test conditions, and source)
Source value
54.808
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
legal_research
Task label
Legal Research
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:legal_research
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 1881.045. Standard error: 3.459. Cost per task in USD: 6.44908. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Full Index (Show measurement, test conditions, and source)
Source value
69.237
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
overall
Task label
Full Index
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:overall
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 2697.953. Standard error: 1.09. Cost per task in USD: 13.372731. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Tax Agent Bench (Show measurement, test conditions, and source)
Source value
44.56
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
tax_agent_bench
Task label
Tax Agent Bench
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:tax_agent_bench
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 985.634. Standard error: 3.587. Cost per task in USD: 3.473645. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Terminal-Bench 4.0 (Show measurement, test conditions, and source)
Source value
57.576
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
terminal_bench_4
Task label
Terminal-Bench 4.0
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:terminal_bench_4
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 3214.793. Standard error: 2.314. Cost per task in USD: 17.63994. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI
Vals Router Index: Vibe Code Bench (Show measurement, test conditions, and source)
Source value
91.906
Metric
accuracy
Unit
%
Benchmark version
1.0
Category
aggregate
Task
vibe_code_bench
Task label
Vibe Code Bench
Direction
Higher is better
Comparison cohort
vals-vals-router-index:1.0:vibe_code_bench
Source type
independent-evaluator
Evaluator
Vals AI
Status
active
Retrieved at
2026-11-07
Methodology
Source model ID: google/gemini-4-argon. Source model name: Gemini 4 Argon. Evaluator: Vals AI. Dataset type: private. Leaderboard date: 2026-11-03. Latency in seconds: 2247.675. Standard error: 1.899. Cost per task in USD: 8.10797. Source ID: vals-ai. Evidence capture ID: vals-vals_router_index-2026-11-07-cefd7c7f7b1e. Source payload hash: cefd7c7f7b1e3bcfee99fdc0c60bdd10976b31ae92748b72d7be3eb561cdedc0. Parser version: vals-astro-v4. Imported at: 2026-11-07
Source Vals AI

Page 1 of 82

Gemini 4 Argon compared

The most popular comparisons

  1. 1Claude Opus 5.5 vs. Gemini 4 Argon
  2. 2Claude Fable 5.2 vs. Gemini 4 Argon
  3. 3GPT-6 Astra vs. Gemini 4 Argon
  4. 4Claude Sonnet 4.5 vs. Gemini 4 Argon
  5. 5GPT-5.6 Sol vs. Gemini 4 Argon
  6. 6GPT-6.1 Terra vs. Gemini 4 Argon
  7. 7Gemini 4 Argon vs. DeepSeek V4.1 Flash 1017
  8. 8Claude Sonnet 5 vs. Gemini 4 Argon
  9. 9GPT-6 Luna vs. Gemini 4 Argon
Show all 59 comparisons
  1. 10Claude Opus 4.6 vs. Gemini 4 Argon
  2. 11Gemini 4 Argon vs. Kimi K3.1 Pro
  3. 12Gemini 4 Argon vs. DeepSeek-V4-Flash
  4. 13Gemini 4 Argon vs. Qwen 3.8 27B
  5. 14Gemini 4 Argon vs. Gemini 3.6 Flash
  6. 15Gemini 4 Argon vs. GLM-5.3
  7. 16Gemini 4 Argon vs. Grok 4.5
  8. 17GPT-6.1 Sol vs. Gemini 4 Argon
  9. 18Claude Fable 5 vs. Gemini 4 Argon
  10. 19Gemini 4 Argon vs. Kimi K3
  11. 20Claude Sonnet 5.5 vs. Gemini 4 Argon
  12. 21GPT-5.6 Luna vs. Gemini 4 Argon
  13. 22Claude Haiku 5.5 vs. Gemini 4 Argon
  14. 23Claude Opus 4.5 vs. Gemini 4 Argon
  15. 24Gemini 4 Argon vs. DeepSeek-V4.1-Flash
  16. 25Gemini 4 Argon vs. DeepSeek-V4-Pro
  17. 26Claude Opus 5 vs. Gemini 4 Argon
  18. 27Gemini 4 Argon vs. Gemini 3.5 Flash-Lite
  19. 28Gemini 4 Argon vs. Qwen 3.8 Max 0902
  20. 29Claude Sonnet 4.6 vs. Gemini 4 Argon
  21. 30Gemini 4 Argon vs. GLM-5.3-Flash
  22. 31GPT-5.6 Terra vs. Gemini 4 Argon
  23. 32Claude Opus 4.8 vs. Gemini 4 Argon
  24. 33Gemini 4 Argon vs. Gemini 3.1 Pro Preview
  25. 34Gemini 4 Argon vs. GLM-5.2
  26. 35Claude Haiku 4.5 vs. Gemini 4 Argon
  27. 36Gemini 4 Argon vs. Grok 4.7
  28. 37Gemini 4 Argon vs. Kimi K2.6
  29. 38Claude Opus 4.7 vs. Gemini 4 Argon
  30. 39Claude Fable 5.1 vs. Gemini 4 Argon
  31. 40Gemini 4 Argon vs. Muse Wonder
  32. 41GPT-6.2 Astra vs. Gemini 4 Argon
  33. 42GPT-6.2 Sol vs. Gemini 4 Argon
  34. 43GPT-6 Sol vs. Gemini 4 Argon
  35. 44GPT-5.5 vs. Gemini 4 Argon
  36. 45GPT-5.4 vs. Gemini 4 Argon
  37. 46GPT-5.4 mini vs. Gemini 4 Argon
  38. 47Gemini 4 Argon vs. Gemini 3.8 Flash
  39. 48Gemini 4 Argon vs. Gemini 3.7 Flash
  40. 49Gemini 4 Argon vs. Gemini 3.5 Flash
  41. 50Gemini 4 Argon vs. Muse Wonder Full
  42. 51Gemini 4 Argon vs. Muse Spark 1.3
  43. 52Gemini 4 Argon vs. Grok 4.6
  44. 53Gemini 4 Argon vs. Grok 4.3
  45. 54Gemini 4 Argon vs. GLM-5.1
  46. 55Gemini 4 Argon vs. Mistral Large 4
  47. 56Gemini 4 Argon vs. Mistral Large 3
  48. 57Gemini 4 Argon vs. Mistral Medium 3.5
  49. 58Gemini 4 Argon vs. MiMo-V2.5
  50. 59Gemini 4 Argon vs. MiMo-V2.5-Pro

Alternatives to Gemini 4 Argon

Stronger, cheaper, open and faster models compared with Gemini 4 Argon. Every card names the values it rests on, and a model appears only once.

Gemini 4 Argon ranks 6 of 95 in the overall index and costs $4 per million tokens.

  • With open weights

    Kimi K3.1 Pro

    The best model with open weights, weaker: 112.6 vs. 121.4 in the overall index.

    Compared with Gemini 4 Argon: 8.9 points fewer in the overall index, the same price, 119 % faster.

    Overall index
    112.6
    Price (3:1)
    $4
    Speed
    78

    Rank 13 · Middle group

    Head to head: Gemini 4 Argon vs. Kimi K3.1 ProOpen the profile

  • Similar strength, faster

    Claude Sonnet 5.5

    Level in the overall index (121.8 vs. 121.4) and faster: 107 instead of 36 tokens per second.

    Compared with Gemini 4 Argon: 0.4 points more in the overall index, the same price, 201 % faster.

    Overall index
    121.8
    Price (3:1)
    $4
    Speed
    107

    Rank 5 · Upper group

    Head to head: Gemini 4 Argon vs. Claude Sonnet 5.5Open the profile

  • The successor

    Gemini 4 Carbon

    The successor to Gemini 4 Argon. Overall index 127.9 vs. 121.4.

    Compared with Gemini 4 Argon: 6.5 points more in the overall index, the same price, 10 % faster.

    Overall index
    127.9
    Price (3:1)
    $4
    Speed
    39

    Rank 2 · Top group

    Head to head: Gemini 4 Argon vs. Gemini 4 CarbonOpen the profile

Where the alternatives lead Gemini 4 Argon

Only areas with a clear lead in the benchmarks. Areas without a clear difference are not listed.

  • Kimi K3.1 Pro

    Clearly ahead in
    none

    Clearly behind in
    Coding, General, Finance, Legal

  • Claude Sonnet 5.5

    Clearly ahead in
    Agents

    Clearly behind in
    none

  • Gemini 4 Carbon

    Clearly ahead in
    Agents

    Clearly behind in
    none

The models around Gemini 4 Argon

The nearest generally available models by overall index, with rank, price and speed. Models with limited access and retired models are left out.

Price in dollars per million tokens, mixed from 3 parts input and 1 part output. Speed in tokens per second.

Cost for your load

1,000 requests with 2,000 input and 500 output tokens cost $9.00 with Gemini 4 Argon.

Compare Gemini 4 Argon with the models next to it in the ranking

All specifications

Input and output

Inputs
Text, image, video, audio, and PDF files
Source: Inputs
Outputs
Text
Source: Outputs
Context window
1,048,576 tokens shared across input, thinking, and output
Source: Context window
Maximum output
65,536 tokens synchronous; up to 1,048,576 with long decode minus input
Source: Maximum output

Watermarking and provenance

AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.

AI watermarking
Subject to limitations. In a study published in October 2026, ETH Zurich found patterns consistent with SynthID Text in Gemini API responses. The study did not verify every individual output or delivery channel.
Source: AI watermarking
Watermarking technique
SynthID Text
Source: Watermarking technique
Labelling reviewed
Nov 7, 2026
Source: Labelling reviewed

Licence and access

Weights
Proprietary
Source: Weights
Announcement
September 30, 2026
Source: Announcement
General availability
October 13, 2026
Source: General availability
Access
Generally available through the Gemini API and Google AI Studio; the Gemini API free tier is unavailable.
Source: Access
Shutdown
None announced
no source

Frequently asked questions about Gemini 4 Argon

How good is Gemini 4 Argon in benchmarks?

Good: Rank 6 of 95 in the overall index.

Overall index (Gradually Meta-Score v6) from 48 benchmark results · data through Nov 19, 2026

IntelligenceOverall index in points, 100 = reference group average · higher is better
Claude Opus 5.5125.1
Claude Fable 5.1124.1
Claude Sonnet 5.5121.8
Gemini 4 ArgonRank 6121.4
GPT-6 Astra119.6
Muse Wonder Full117.9
GPT-6.1 Sol117.4

What is the Overall index (Gradually Meta-Score v6)?

A point value that combines many independent benchmarks into one number. 100 points is the average of a fixed group of 29 models, and a model 15 points ahead does better in roughly three of four benchmark comparisons. Gemini 4 Argon scores 121.4 points, which is rank 6 of 95.

Own analysis by gradually.ai, method benchmark-meta-score-v6 · for Gemini 4 Argon from 48 test series

How much does Gemini 4 Argon cost?

Through the API $2 per 1M input tokens and $10 output. That is fairly expensive.

PriceDollars per 1M tokens, input and output blended · lower is better
Gemini 4 ArgonRank 76 of 102$4
Claude Sonnet 5.5$4
GPT-6.1 Sol$4
Gemini 4 Carbon$4
Claude Opus 5.5$8
GPT-6 Astra$20
Claude Fable 5.1$20

How large is the context window of Gemini 4 Argon?

1.05 M tokens. Separate limits may apply to input and output.

Context windowContext limit in tokens · higher is better
GPT-6 Astra1.05M
GPT-6.1 Sol1.05M
Gemini 4 ArgonRank 12 of 1111.05M
Claude Sonnet 5.51M
Claude Fable 5.11M
Muse Wonder Full1M
Claude Opus 5.51M

How fast is Gemini 4 Argon?

36 tokens per second. That is slower than a typical model at 70.

Measured by OpenRouter on Nov 20, 2026. Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token.

SpeedTokens per second · higher is better
Claude Sonnet 5.5107
Claude Opus 5.579
Claude Fable 5.165
GPT-6 Astra56
GPT-6.1 Sol46
Gemini 4 Carbon39
Gemini 4 ArgonRank 92 of 9936

Can I run Gemini 4 Argon locally?

No. Google DeepMind does not publish the weights.

Stated by Google DeepMind

Open weights
no
License
n/a
Parameters
n/a
active per token
n/a

Where can I use Gemini 4 Argon?

At Google DeepMind with a subscription from $19.99 a month. Including other providers that makes 30 plans in Germany, plus the API.

Subscription comparison, prices in Germany incl. VAT

At Google DeepMind

3 plans, from $19.99 a month (Google AI Pro).

Through other providers

27 plans from 8 providers, from $6 a month (ElevenLabs Starter), with limited credits.

Through the API

$2 input, $10 output per 1M tokens.

Sources and data date

Show all 114 sources

Benchmark sources at a glance

Which independent evaluators tested this model. This is coverage, not a quality ranking.

EvaluatorTest seriesComposite indicesRecorded active measurements
Artificial Analysis12014
Vals AI345305
LMArena2033
LiveBench7147
Epoch AI6010
ARC Prize006
Scale AI Evals0013

Without vendor claims and historical or disputed results; those stay labelled in the single values. 30 measurements are not yet assigned to a test series.

TypeEvidence and data date
Speed

OpenRouter. Median across providers on OpenRouter from the readings of the last 24 hours (each covering the last 30 minutes), excluding time to first token. Retrieved November 20, 2026.

Price

Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs.

Input

Documented modalities.

Output

Documented modalities.

Research statusResearch date October 4, 2026. 61 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication.
Not published

MultiChallenge model-specific run settings and measurement date

Note

The page publishes leaderboard methodology but does not identify this row's measurement date or model-side generation settings.

Not published

PRBench row subset and model-specific run settings

Note

The current leaderboard page does not identify whether these rows use the full or hard subset, or publish the model-specific generation settings or measurement date.

Additional sourceRelease announcement (retrieved October 25, 2026) · Google Gemini API release notes · Editorial description reviewed October 25, 2026
Additional sourceKnowledge cutoff data (retrieved October 17, 2026; October 21, 2026)
Additional sourceGoogle Gemini 4 Argon announcement (retrieved October 4, 2026; October 25, 2026) · Google Gemini 4 Argon announcement · Editorial description reviewed October 25, 2026
Additional sourceGoogle Gemini API overview (retrieved October 25, 2026) · Google Gemini API latest-model documentation · Editorial description reviewed October 25, 2026
Additional sourceArtificial Analysis (retrieved November 2, 2026; November 7, 2026)12345678
Additional sourceArtificial Analysis (retrieved October 29, 2026; November 2, 2026; November 7, 2026; November 10, 2026)12
Additional sourceArtificial Analysis (retrieved November 2, 2026)12
Additional sourceArtificial Analysis (retrieved November 7, 2026)123
Additional sourceLMArena (retrieved October 30, 2026)12
Additional sourceLMArena (retrieved November 13, 2026)12
Additional sourceLMArena (retrieved October 28, 2026)12
Additional sourceLMArena (retrieved November 9, 2026)12
Additional sourceLMArena (retrieved November 19, 2026)12
Additional sourceLMArena (retrieved November 10, 2026)12
Additional sourceLMArena (retrieved November 16, 2026)
Additional sourceVals AI (retrieved November 10, 2026)
Additional sourceVals AI (retrieved November 19, 2026)
Additional sourceVals AI (retrieved November 7, 2026)1234567891011121314151617181920212223242526
Additional sourceVals AI (retrieved November 7, 2026; November 10, 2026)123
Additional sourceVals AI (retrieved November 13, 2026)
Additional sourceVals AI (retrieved October 16, 2026; November 7, 2026)123
Additional sourceVals AI (retrieved October 16, 2026; October 31, 2026; November 7, 2026)
Additional sourceVals AI (retrieved October 30, 2026; October 31, 2026)
Additional sourceVals AI (retrieved October 16, 2026; October 17, 2026; November 7, 2026)
Additional sourceVals AI (retrieved October 16, 2026; November 7, 2026; November 10, 2026)
Additional sourceVals AI (retrieved October 28, 2026; November 7, 2026)
Additional sourceLiveBench (retrieved October 16, 2026)1 (unavailable on November 2, 2026) https://livebench.ai/table_2026_09_02.csv2 (unavailable on November 2, 2026) https://livebench.ai/breakdown_2026_09_02.json
Additional sourcePlotly, Gemini 4 Argon Benchmark Results for Data Analysis (retrieved October 17, 2026)
Additional sourceTextQL Argo-Bench official leaderboard (retrieved October 17, 2026)
Additional sourceARC Prize verified results (retrieved October 18, 2026; October 19, 2026)
Additional sourceARC Prize v2 official leaderboard data (retrieved October 18, 2026)
Additional sourceEpoch AI benchmark run dataset snapshot (retrieved October 19, 2026; October 28, 2026)
Additional sourceRoboflow Vision Evals (retrieved October 19, 2026)
Additional sourceLiveBench (retrieved October 28, 2026)12
Additional sourceScale Labs MultiChallenge (retrieved October 28, 2026)
Additional sourceScale AI (retrieved October 28, 2026)12
Additional sourceGoogle on Kaggle (retrieved October 28, 2026)1234
Additional sourceDuelLab GameBench 2 model results (retrieved October 28, 2026; November 2, 2026; November 5, 2026; November 9, 2026; November 11, 2026; November 12, 2026; November 15, 2026)
Additional sourceParallel Search Capability Leaderboard (retrieved October 28, 2026; November 5, 2026)
Additional sourceTowards AI ToneBench (retrieved October 28, 2026; October 30, 2026; November 10, 2026; November 11, 2026)12
Additional sourceRoboflow Vision Evals (retrieved October 28, 2026)
Additional sourceGoogle DeepMind on Kaggle (retrieved October 28, 2026)
Additional sourceMathBench.ai (retrieved October 28, 2026)
Additional sourceScale AI Labs (retrieved October 28, 2026)
Additional sourceScale AI HSS (retrieved October 28, 2026)
Additional sourceScale AI DrugDiscoveryBench (retrieved October 28, 2026)
Additional sourceScale AI HiL-Bench (retrieved October 28, 2026)
Additional sourceScale AI AudioMultiChallenge (retrieved October 28, 2026)
Additional sourceScale AI RLI (retrieved October 28, 2026)
Additional sourceScale SWE-Bench Pro V2 public leaderboard (retrieved October 28, 2026)
Additional sourceParallel Search Capability Leaderboard (retrieved November 5, 2026)
Additional sourceScale AI DistressBench (retrieved November 5, 2026)
Additional sourceScale AI (retrieved November 8, 2026)
Additional sourceLiveBench (retrieved November 10, 2026)12
Additional sourceTerminal-Bench (retrieved November 14, 2026)
Additional sourceETH Zurich study of Gemini API text (retrieved November 7, 2026)
Additional sourceGoogle: Introducing Gemini 4 Carbon (2026-10-28)