Gemini 4 Argon

GoogleKey gemini-4-argonAlso reported as google-antigravity/gemini-4-argon, cursor/gemini-4-argon, openrouter/google/gemini-4-argon, google/gemini-4-argon, google/models/gemini-4-argon, ai-studio/gemini-4-argon, devin/gemini-4-argon, opencode-go/gemini-4-argon

Rank
#27
by active users
Active users
82
Up 77 from 5 active users in the previous 30 days. vs previous 30 days
Adoption
8.6%
of 958 accounts
Tokens
82B
cache included
Est. cost
$18.3K
at public list prices
First seen
Oct 10, 2026
in any submission

Oct 14 – Nov 12, 2026 (UTC)958 accounts100% of tokens from an identified modelData through Nov 12, 2026, 00:36 UTCHow we count

Daily active users

Distinct accounts that used this model on each UTC day.

2 active users on Nov 12, 2026

Daily active users for Gemini 4 Argon, Oct 14, 2026 to Nov 12, 2026. Peak 26. Use the left and right arrow keys to read each day.30150
Show daily values
Daily active users and tokens for Gemini 4 Argon
Date (UTC)Active usersTokens
Oct 14, 2026206.4B
Oct 15, 2026213.3B
Oct 16, 2026263.1B
Oct 17, 2026215.8B
Oct 18, 2026217.5B
Oct 19, 2026206B
Oct 20, 2026183.8B
Oct 21, 2026192.6B
Oct 22, 2026181.2B
Oct 23, 202612772.8M
Oct 24, 2026121.8B
Oct 25, 2026165B
Oct 26, 2026204.2B
Oct 27, 2026202.5B
Oct 28, 2026184.4B
Oct 29, 2026201.6B
Oct 30, 202613695.4M
Oct 31, 2026152.3B
Nov 1, 2026152.4B
Nov 2, 2026202.5B
Nov 3, 2026182.4B
Nov 4, 2026121.5B
Nov 5, 2026122B
Nov 6, 2026121.2B
Nov 7, 20269683.6M
Nov 8, 202691.8B
Nov 9, 2026132.7B
Nov 10, 2026121.1B
Nov 11, 20267387.3M
Nov 12, 2026250.9M

Token mix

How the 82B tokens split. Cache reads are usually most of an agent's tokens and are priced far below input.

  • Input7.1%
  • Output0.3%
  • Cache read92%
  • Cache write0%
  • Reasoning0.3%

Agents that carry Gemini 4 Argon

Each agent's share of this model's tokens, and how many accounts used the model through it.

  1. Antigravity CLI56% of this model's tokens26active users
  2. Oh My Pi25% of this model's tokens18active users
  3. Antigravity IDE Extension5.7% of this model's tokens7active users
  4. Antigravity4.7% of this model's tokens4active users
  5. Pi3.8% of this model's tokens11active users
  6. Senpi (OmO Native)2.5% of this model's tokens10active users
  7. OpenCode1.1% of this model's tokens12active users
  8. Hermes Agent0.3% of this model's tokens3active users
  9. Droid0.2% of this model's tokens1active users
  10. Aside0.1% of this model's tokens2active users
  11. Codex CLI0.1% of this model's tokens4active users
  12. Claude Code<0.1% of this model's tokens6active users
  13. Gajae Code<0.1% of this model's tokens1active users
  14. DeepSeek Harness<0.1% of this model's tokens3active users
  15. Copilot<0.1% of this model's tokens1active users
  16. ZCode<0.1% of this model's tokens1active users
  17. Cline<0.1% of this model's tokens1active users
  18. Zed Agent<0.1% of this model's tokens1active users
  19. WorkBuddy<0.1% of this model's tokens1active users

Benchmarks

The best result for Gemini 4 Argon in each benchmark Tokscale tracks. Each benchmark has its own unit, run conditions and source, so compare a score only with other models on the same benchmark. Rank is among the models Tokscale matched in that benchmark.

All benchmarks
  • Epoch Capabilities IndexOverall · Epoch AI
    166.4
    Rank #5 of 160 matched modelsNo conditions reportedModel released Sep 30, 2026Source: epoch.ai (opens in a new tab)
  • CursorBenchAgentic coding · Cursor, collected by Epoch AI
    52.9%
    Rank #3 of 18 matched modelsHighModel released Sep 30, 2026Source: cursor.com (opens in a new tab)
    Show 3 more results with other run conditionsHide the other results
    • Listed as Gemini 4 Argon Medium
      52.8%
      MediumModel released Sep 30, 2026Source: cursor.com (opens in a new tab)
    • Listed as Gemini 4 Argon Low
      50.6%
      LowModel released Sep 30, 2026Source: cursor.com (opens in a new tab)
    • Listed as Gemini 4 Argon Minimal
      48.8%
      MinimalModel released Sep 30, 2026Source: cursor.com (opens in a new tab)
  • FrontierSWEAgentic coding · FrontierSWE, collected by Epoch AI
    55.0%
    Rank #5 of 21 matched modelsHigh · proximusModel released Sep 30, 2026Source: frontierswe.com (opens in a new tab)
  • WebDev ArenaWeb development · LMArena
    1,678
    Rank #9 of 110 matched modelsHighPublished Oct 19, 2026Source: huggingface.co (opens in a new tab)
  • Agent ArenaAgents · LMArena
    +0.12
    Rank #6 of 48 matched modelsHigh · Arena Agent ModePublished Oct 19, 2026Source: huggingface.co (opens in a new tab)
  • GPQA DiamondReasoning · Epoch AI
    96.4%
    Rank #1 of 151 matched modelsHighRun date Oct 13, 2026Source: epoch.ai (opens in a new tab)

    Near its ceiling: the top score is 96.4%, and Epoch AI estimates the reachable maximum at 97%. Differences between top models are small.

  • FrontierMath Tiers 1–3 (v2)Reasoning · Epoch AI
    86.0%
    Rank #10 of 86 matched modelsHighRun date Oct 13, 2026Source: epoch.ai (opens in a new tab)

    Near its ceiling: the top score is 93.7%, and Epoch AI estimates the reachable maximum at 95%. Differences between top models are small.

  • SimpleQA VerifiedKnowledge · Epoch AI
    76.0%
    Rank #1 of 80 matched modelsHighRun date Oct 13, 2026Source: epoch.ai (opens in a new tab)

Not evaluated: Humanity's Last Exam, Terminal-Bench 4.0, DeepSWE. A missing result is not a score of zero.

Pricing

List prices from models.dev in US dollars per 1M tokens, retrieved Oct 21, 2026. Google is the lab's own API and the reference price for the estimates below.

Price at real agent usageEstimate at list prices

What 1M tokens cost at the reference list price when they split into input, output and cache the way Tokscale users' agents used them. It is not what anyone paid.

At this model's mix
$0.30
Est. per 1M tokens
92% cache reads
At the community mix
$0.20
Est. per 1M tokens
96% cache reads
3:1 input/output blend
$4
Per 1M tokens at list prices
No cache

Cache reads were 92% of Gemini 4 Argon tokens in the last 30 days and 96% of all tokens. The 3:1 blend, (3 × input + output) ÷ 4, is the usual comparison and ignores cache.

Assumptions

  • Google list prices per 1M tokens: input $2, output $10, cache read $0.10.
  • This model's mix is its own input, output, cache and reasoning tokens in the last 30 days.
  • The community mix is the tokens of every identified model in the last 30 days.
  • This is an estimate at list prices, not what anyone paid.
  • Reasoning tokens are priced as output.
  • Long-context tiers, batch pricing and discounts are not applied.
  • Token mixes come from public Tokscale submissions. Hidden accounts are excluded.
  • No cache-write price is listed, so cache writes are priced as input.

List prices by provider

  • GoogleReferencegemini-4-argon
    $2 input per 1M tokens$10 output per 1M tokens$0.10 cache read per 1M tokensCache write not listed1.05M tokens of context
  • Vertexgemini-4-argon
    $2 input per 1M tokens$10 output per 1M tokens$0.10 cache read per 1M tokensCache write not listed1.05M tokens of context
  • OpenRoutergoogle/gemini-4-argon
    $2 input per 1M tokens$10 output per 1M tokens$0.10 cache read per 1M tokens$0.19 cache write per 1M tokens1.05M tokens of context
  • Vercel AI Gatewaygoogle/gemini-4-argon
    $2 input per 1M tokens$10 output per 1M tokens$0.10 cache read per 1M tokensCache write not listed1M tokens of context

Specs

From models.dev (MIT), retrieved Oct 21, 2026, as the lab or the first listed provider reports them. A dash means models.dev does not list the value.

Context window
1.05M tokens
Max output
33K tokens
Input
Text, Image, Video, Audio, PDF
Output
Text
Knowledge cutoff
Mar 2026
Release date
Oct 13, 2026
Open weights
No
Reasoning
Yes
Tool calling
Yes

Top public profiles

Public profiles that used Gemini 4 Argon most in this period, by tokens. Hidden accounts are never listed.

  1. 1Jeff Emanuel@Dicklesworthstone21.9Btokens$5.6Kest. cost15active days$5.6K est. cost · 15 active days
  2. 2Brad Lishman@bradlishman12.4Btokens—est. cost25active daysno price · 25 active days
  3. 3WONJUN SEO@ownuun7.4Btokens$2.4Kest. cost16active days$2.4K est. cost · 16 active days
  4. 4KIM SUNGMIN@trinity-mathslab5.4Btokens$1.9Kest. cost22active days$1.9K est. cost · 22 active days
  5. 5Joosung Park@hyperlapse1224.9Btokens$719est. cost18active days$719 est. cost · 18 active days
  6. 6Muhammad Abdullah@iAbdullahMughal4.8Btokens$1.5Kest. cost18active days$1.5K est. cost · 18 active days
  7. 7Hoang-Long Do@hl23114.4Btokens$1.4Kest. cost21active days$1.4K est. cost · 21 active days
  8. 8Woonggi Min@minpeter4.2Btokens$6est. cost14active days$6 est. cost · 14 active days
  9. 9Nehr@AgileInnov8tor4.1Btokens$965est. cost23active days$965 est. cost · 23 active days
  10. 10PROLL@DEV-PROLL1.6Btokens—est. cost17active daysno price · 17 active days

More from Google

Other Google models in the same period, by active users.

  1. Gemini 3.8 Flash N14% adoption138active users
  2. Gemini 4 Argon Trial7.7% adoption74active users
  3. Gemini 3.8 Flash6.9% adoption66active users
  4. Gemini 3.8 Flash Tiered3.7% adoption35active users
  5. Gemini 4 Argon High1.8% adoption17active users
  6. Gemini 3.8 Flash High1.6% adoption15active users
All models

Sources and attribution

Each score keeps its own source, unit and run conditions. Tokscale does not combine benchmarks into one score, and a missing score means the model was not evaluated, not that it scored zero.

  • Epoch AIEpoch Capabilities Index, Epoch's own runs and the external results it collects.CC BY 4.0 · Retrieved Oct 21, 2026 · Data through Oct 15, 2026Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from https://epoch.ai/benchmarks.
  • LMArenaWebDev Arena and Agent Arena leaderboards.CC BY 4.0 · Retrieved Oct 21, 2026 · Data through Oct 19, 2026LMArena leaderboard dataset (lmarena-ai/leaderboard-dataset), Hugging Face.
  • models.devSpecs and list prices per provider.MIT · Retrieved Oct 21, 2026models.dev, an open catalog of AI models (github.com/anomalyco/models.dev).
  • TokscaleActive users, tokens and token mix from public Tokscale submissions. Hidden accounts are excluded.How we count