Gemini 4 Argon
Gemini 4 Argon is at the frontier in factuality; strong in reasoning; and capable in long context, agentic tasks, multimodal tasks, and math. Too few results yet to rate coding, safety, multilingual tasks, or instruction following.
Capability profile
Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.
Too few results yet to rate Coding, Safety, Multilingual or Instruction Following.
Price
$2.00input$10.00outputper million tokens
From Artificial Analysis · 3 providers tracked · All prices
Evidence
94results on74benchmarks
- 25 independently verified
- 26 aggregator
- 43 vendor-reported
From 20 sources · latest Nov 16, 2026 · How verification works
API features
Tool callingStructured outputsReasoning
As listed by OpenRouter
Gemini 4 Argon benchmark results
94 results on 74 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.
0.2% behind the leader4 of 4 ranked benchmarks measured
- AA-Omniscience · Non-hallucination84.93Oct 16, 2026omniscienceNonHallucination
- 12.10Oct 17, 2026
- 76.00Oct 14, 2026
- AA-Omniscience · Accuracy49.90Oct 16, 2026omniscienceAccuracy
9.6% behind the leader4 of 6 ranked benchmarks measured
- 91.70Nov 11, 2026
- 91.25Oct 16, 2026
- Humanity's Last Exam57.09Nov 16, 2026aa_hle
- 27.14Nov 16, 2026
Show 9 more reasoning resultsHide 9 reasoning results
- 42.08Oct 16, 2026
- 87.08Oct 16, 2026
- 97.46Oct 16, 2026
- 97.34Oct 16, 2026
- 97.30Oct 16, 2026
- 49.79Oct 16, 2026
- 46.27Oct 16, 2026
- 23.14Nov 16, 2026
- Humanity's Last Exam49.91Nov 16, 2026aa_hle
18.0% behind the leader1 of 3 ranked benchmarks measured
- 79.67Nov 16, 2026
Show 1 more long context resultHide 1 long context result
- 78.33Nov 16, 2026
18.3% behind the leader4 of 7 ranked benchmarks measured
- 90.70Nov 16, 2026
- Terminal-Bench 4.057.07Nov 16, 2026
- 56.20Oct 16, 2026
Show 1 more agentic resultHide 1 agentic result
- Terminal-Bench 4.055.05Nov 16, 2026
21.2% behind the leader1 of 6 ranked benchmarks measured
- MMMU-Pro88.32Nov 16, 2026aa_mmmu_pro
23.7% behind the leader2 of 5 ranked benchmarks measured
- 85.96Oct 15, 2026
- 78.05Oct 15, 2026
0 of 10 ranked benchmarks measured
0 of 3 ranked benchmarks measured
- 78.70Nov 16, 2026
More results
Benchmarks outside the capability baskets. They are not ranked against a leader.
- 62.00Nov 16, 2026
- AA Intelligence48.00Nov 14, 2026Artificial Analysis Intelligence Index
- vectara_avg_summary_length178.20Oct 17, 2026Average Summary Length (Words)
- vectara_answer_rate99.10Oct 17, 2026Answer Rate
- vectara_factual_consistency87.90Oct 17, 2026Factual Consistency Rate
- 76.50Oct 16, 2026
Show 58 more resultsHide 58 results
- AA Intelligence50.00Sep 30, 2026Artificial Analysis Intelligence Index
- AA Intelligence52.56Oct 16, 2026aa_intelligence_index
- 42.35Oct 16, 2026
- 39.50Oct 1, 2026
- Agentic coding (Terminal-Bench 4.0 (AA))57.10Oct 29, 2026
- 57.40Oct 14, 2026
- 39.50Nov 16, 2026
- 98.50Oct 16, 2026
- 1525Oct 15, 2026
- 51.30Oct 1, 2026
- Computer use (Agent's Last Exam)39.50Oct 14, 2026
- 69.00Oct 29, 2026
- 69.20Oct 14, 2026
- 74.70Oct 29, 2026
- 97.80Nov 3, 2026
- 97.80Nov 2, 2026
- 83.70Nov 3, 2026
- 83.70Nov 2, 2026
- CritPt v2 Hard52.50Nov 16, 2026
- CritPt v2 Hard53.50Nov 16, 2026
- 68.00Oct 14, 2026
- 68.00Oct 29, 2026
- 77.90Nov 16, 2026
- 87.00Nov 6, 2026
- 87.00Nov 2, 2026
- GDP (Surge AI)21.80Nov 16, 2026gdp-pdf
- GDP (Surge AI)25.60Nov 16, 2026gdp-pdf
- 50.00Oct 29, 2026
- 69.20Oct 29, 2026
- GraphWalks (256k to 1M, BFS (F1))84.20Oct 1, 2026
- GraphWalks (Up to 128k, BFS (F1))99.70Oct 1, 2026
- 49.05Nov 16, 2026
- HealthBench Hard45.40Nov 2, 2026
- HealthBench Pro, with Search grounding76.70Nov 2, 2026
- HealthBench Pro, with Search grounding (length-adjusted)61.40Nov 2, 2026
- HealthBench Pro, without Search grounding68.80Nov 2, 2026
- HealthBench Pro, without Search grounding (length-adjusted)56.50Nov 2, 2026
- 55.05Nov 16, 2026
- HLE-Diamond56.70Nov 16, 2026
- HLE-Diamond42.50Nov 16, 2026
- 51.30Oct 14, 2026
- 68.90Oct 14, 2026
- 91.70Nov 16, 2026
- 91.70Oct 14, 2026
- 69.20Nov 16, 2026
- OSWorld-2.0 (Offline subset)69.20Oct 1, 2026
- 45.30Nov 16, 2026
- 57.00Nov 2, 2026
- 76.00Nov 16, 2026
- 76.00Oct 14, 2026
- Science and math (Terminal-Bench Science 0.1)57.60Oct 14, 2026
- Science and math (Terminal-Bench-Science 0.1 (AA))52.90Oct 29, 2026
- Terminal-Bench 4.057.40Nov 16, 2026
- Terminal-Bench-Science 0.152.86Nov 16, 2026
- Terminal-Bench-Science 0.148.57Nov 16, 2026
- Terminal-Bench-Science 0.157.60Nov 16, 2026
- 68.90Nov 16, 2026
- 1.00Nov 2, 2026
Gemini 4 Argon: common questions
Who makes Gemini 4 Argon?
Gemini 4 Argon is made by Google.
When was Gemini 4 Argon released?
Gemini 4 Argon was released on Sep 30, 2026, according to Google's own announcement.
What is Gemini 4 Argon good at?
Gemini 4 Argon is at the frontier in factuality; strong in reasoning; and capable in long context, agentic tasks, multimodal tasks, and math. Too few results yet to rate coding, safety, multilingual tasks, or instruction following.
How much does Gemini 4 Argon cost?
Gemini 4 Argon costs $2.00 per million input tokens and $10.00 per million output tokens, according to Artificial Analysis. We track its price at 3 providers. At a mix of three input tokens to one output token, it costs more than 82% of the 332 priced models we track.
How many benchmarks has Gemini 4 Argon been tested on?
We track 94 results for Gemini 4 Argon on 74 benchmarks from 20 sources, 25 of them independently verified. The latest was recorded on Nov 16, 2026.
Which API features does Gemini 4 Argon support?
OpenRouter lists tool calling, structured outputs, and reasoning for Gemini 4 Argon.
About this record
Where Gemini 4 Argon's numbers come from, and every name it appears under.
- Tracked since
- Sep 30, 2026
- Newest source mention
- Nov 14, 2026
Where the results come from
Verification: 94 scores · 25 independently verified · 26 aggregator-attributed · 43 vendor-reported. How these tiers are assigned
From 20 sources on 12 sites. Artificial Analysis supplies 28 of them; the 25 independently verified results come from 8 sites. Bars are coloured by trust tier.
- artificialanalysis.ai28
- deepmind.google21
- arcprize.org10
- api.llm-stats.com9
- storage.googleapis.com9
- blog.google4
- raw.githubusercontent.com4
- epoch.ai3
- labs.scale.com3
- datasets-server.huggingface.co1
- lmarena.ai1
- simple-bench.com1
Also known as
How our sources name Gemini 4 Argon at each reasoning setting.
| Setting | Short form | API id |
|---|---|---|
| low | gemini 4 argon (low) | — |
| medium | gemini 4 argon (medium) | gemini-4-argon-medium |
| high | gemini 4 argon (high) | gemini-4-argon-high |