GoogleGemini 4 Argon
Gemini 4 Argon is a Google proprietary reasoning model, released 30 Sept 2026. Its best configuration (high reasoning effort) ranks #2 of 775 on the BenchLeader Index at 71.0 ±2.6, in the top ten. The ± is the point: 37 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in composite (89) and lowest in long context (65). At $4.00 per million tokens blended it is among the most expensive ranked models. Output speed of 83 tokens per second puts it slower than most, with a first answer in 48.5 s. It has been measured at 5 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 19 Oct 2026.
- Blended price
- $4.00/M
- $2.00 in · $10.00 out
- Output speed
- 83 tok/s
- measured by Artificial Analysis
- First answer
- 49 s
- first token 4.23 s
- Context
- 1.0M
- Full answer
- 55 s
- median, reasoning included
- Cost per run
- $1.99
- one full Intelligence Index run
- Released
- 30 Sept 2026
- Overall index71
- Reasoning75
- Coding72
- Agents & tools72
- Maths71
- Knowledge77
- Human preference73
- Long context65
- Composite89
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| minimal | – | – | 40 tok/s | 4.23 s | $0.0038 | |
| low | – | – | 40 tok/s | 4.23 s | $0.0038 | Reasoning 60 |
| medium | – | – | 40 tok/s | 4.23 s | $0.0038 | Coding 62 · Reasoning 72 |
| highbest | 71.0 | #2 | 83 tok/s | 49 s | $0.0038 | Agents & tools 72 · Coding 72 · Composite 89 · Human preference 73 · Knowledge 77 · Long context 65 · Maths 71 · Reasoning 75 |
| not stated | 63.3 | #66 | 40 tok/s | 4.23 s | $0.0038 | Agents & tools 75 · Composite 68 · Knowledge 59 · Maths 78 |
How its index has moved
2 Oct 2026 to 21 Oct 2026- high
- not stated
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | low | medium | high | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| GPQA Diamond | – | – | 96.4%#1 | – | Epoch AI Benchmarking Hub | |
| Arena Hard Prompts | – | – | 1549#1 | – | Arena (LMArena) | |
| LiveBench 2.0 Reasoningnot in indexeffort not stated | – | – | – | 95.1%#1 | LiveBench | |
| Humanity's Last Exam (AA)not in index | – | – | 57.1%#5 | – | Artificial Analysis | |
| ARC-AGI-1 | 76.5%#120 | 98.5%#1 | 98.5%#1 | – | ARC Prize | |
| ARC-AGI-2 | 42.1%#110 | 87.1%#25 | 91.3%#12 | – | ARC Prize | |
| ARC-AGI-3 | 97.5%#6 | 97.3%#7 | 97.3%#7 | – | ARC Prize | |
| CritPt | – | – | 27.1%#33 | – | Artificial Analysis | |
| MysteryMechanismnot in index | – | – | 45.5%#6 | – | Vals AI | |
| Chess Puzzlesnot in index | – | – | 59.0%#10 | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | – | – | 64.0%#5 | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | 62.9%#25 | 62.9%#26 | 63.7%#19 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | 97.6%#5 | 97.3%#9 | 97.3%#9 | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | minimal | low | medium | high | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| SciCode | – | – | – | 61.8%#4 | – | SciCode | |
| FrontierCode | – | – | 47.5%#12 | – | – | Cognition | |
| Arena Coding | – | – | – | 1559#1 | – | Arena (LMArena) | |
| Arena WebDev | – | – | – | 1678#10 | – | Arena (LMArena) | |
| LiveBench 2.0 Codingnot in indexeffort not stated | – | – | – | – | 89.7%#14 | LiveBench | |
| SciCode (AA)not in index | – | – | – | 61.8%#5 | – | Artificial Analysis | |
| IOI | – | – | – | 100.0%#1 | – | Vals AI | |
| Code Migrationnot in index | – | – | – | 68.2%#2 | – | Vals AI | |
| ProgramBenchnot in index | – | – | – | 2.5%#7 | – | Vals AI | |
| Vibe Code Bench 1-100not in index | – | – | – | 21.2%#8 | – | Vals AI | |
| Vibe Code Bench v1.1not in index | – | – | – | 91.9%#2 | – | Vals AI | |
| CursorBenchnot in index | 48.8%#14 | 50.6%#11 | 52.8%#7 | 52.9%#6 | – | Cursor | |
| ALE-Benchnot in index | – | 1705.4#13 | – | 2100.1#8 | – | Epoch AI Benchmarking Hub | |
| FrontierSWEnot in index | – | – | – | 55.0%#5 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | – | – | – | 57.1%#6 | – | Artificial Analysis | |
| FrontierCode 1.1 (AA)not in index | – | – | – | 43.5%#17 | – | Artificial Analysis |
Agents & tools
Maths
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| FrontierMath Tiers 1–3 | 86.0%#10 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | 78.0%#15 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 100.0%#1 | – | Epoch AI Benchmarking Hub | |
| ProofBencheffort not stated | – | 99.0%#4 | Vals AI | |
| Arena Maths | 1529#1 | – | Arena (LMArena) |
Knowledge
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| SimpleQA Verified | 76.0%#1 | – | Epoch AI Benchmarking Hub | |
| LiveBench 2.0 Data Analysisnot in indexeffort not stated | – | 85.2%#1 | LiveBench | |
| LiveBench Researchnot in indexeffort not stated | – | 83.1%#12 | LiveBench | |
| AA-Omniscience | 42.4#9 | – | Artificial Analysis | |
| LegalBench | 88.3%#3 | – | Vals AI | |
| PRBench Financeeffort not stated | – | 49.9%#11 | Scale AI SEAL | |
| PRBench Legaleffort not stated | – | 49.4%#12 | Scale AI SEAL | |
| BioMysteryBenchnot in index | 76.3%#6 | – | Vals AI | |
| Excel Modeling Benchmarknot in index | 75.2%#4 | – | Vals AI | |
| MedCodenot in index | 58.8%#3 | – | Vals AI | |
| MedScribenot in index | 87.4%#17 | – | Vals AI | |
| Public Benefits Benchnot in index | 69.8%#7 | – | Vals AI | |
| EBR-benchnot in index | 64.8%#3 | – | Epoch AI Benchmarking Hub | |
| GDP.pdfnot in index | 28.4%#11 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 49.9%#62 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 84.9%#5 | – | Artificial Analysis |
Instruction following
| Benchmark | high | Source | Trend |
|---|---|---|---|
| Arena Instruction Followingnot in index | 1527#1 | Arena (LMArena) |
Human preference
| Benchmark | high | Source | Trend |
|---|---|---|---|
| Arena Text | 1524#1 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1516#2 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1552#1 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1543#1 | Arena (LMArena) |
Multimodal
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| LiveBench Designnot in indexeffort not stated | – | 73.7%#31 | LiveBench | |
| SAGEnot in index | 53.6%#4 | – | Vals AI | |
| AudioMCnot in index | 58.4%#3 | – | Scale AI SEAL | |
| AudioMC (text output)not in index | 58.4%#3 | – | Scale AI SEAL | |
| Blueprint-Bench 2not in indexeffort not stated | – | 54.5%#1 | Epoch AI Benchmarking Hub | |
| AA-3Dnot in index | 1154#4 | – | Artificial Analysis |
Long context
| Benchmark | high | Source | Trend |
|---|---|---|---|
| AA-LCR | 79.7%#87 | Artificial Analysis |
Composite
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | 166.4#5 | Epoch AI Benchmarking Hub | |
| LiveBench 2.0effort not stated | – | 84.1%#5 | LiveBench | |
| AA Intelligence Index v4.3.2 | 52.6#9 | – | Artificial Analysis | |
| Vals Indexnot in index | 69.2#3 | – | Vals AI |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Google AI Studio | 46 tok/s | 4.70 s | $1.00 | $5.00 | 1.0M | – |
| Google Vertex | 40 tok/s | 3.33 s | $3.60 | $18.00 | 1.0M | – |
| Google AI Studio | 40 tok/s | 6.33 s | $2.00 | $10.00 | 1.0M | – |
| Google Vertex | 39 tok/s | 4.58 s | $2.00 | $10.00 | 1.0M | – |
| Google Vertex | 27 tok/s | 3.95 s | $2.20 | $11.00 | 1.0M | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.100 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0038 | $0.0032 | 52.2 s |
| Summarize a 30-page report | 12,000 / 600 | $0.030 | $0.013 | 55.8 s |
| Code edit | 6,000 / 1,500 | $0.027 | $0.018 | 1.1 min |
| Agentic coding session | 60,000 / 4,000 | $0.160 | $0.074 | 1.6 min |
| Structured extraction | 2,000 / 200 | $0.0060 | $0.0032 | 51.0 s |
More from Google
Compare Gemini 4 Argon with…
- Gemini 4 Argon vs Claude Opus 5.5
- Gemini 4 Argon vs Muse Wonder Full
- Gemini 4 Argon vs GPT-6 Astra
- Gemini 4 Argon vs Claude Fable 5.1
- Gemini 4 Argon vs GPT-6.1 Sol
- Gemini 4 Argon vs Claude Sonnet 5.5
- Gemini 4 Argon vs Claude Fable 5
- Gemini 4 Argon vs Claude Opus 5
- Gemini 4 Argon vs GPT-5.6 Sol
- Gemini 4 Argon vs Muse Wonder
- Gemini 4 Argon vs Muse Spark 1.3
- Gemini 4 Argon vs GPT-6 Sol
- Gemini 4 Argon vs GPT-5.5
- Gemini 4 Argon vs GPT-5.5 Pro
- Gemini 4 Argon vs Kimi K3
- Gemini 4 Argon vs GPT-5.4
- Gemini 4 Argon vs Gemini 3.8 Flash
- Gemini 4 Argon vs Muse Spark
- Gemini 4 Argon vs GPT-5.6 Terra
- Gemini 4 Argon vs MiMo-V2.6-Pro
- Gemini 4 Argon vs Grok 4.6
- Gemini 4 Argon vs Grok 4.7
- Gemini 4 Argon vs GPT-5.3-Codex
- Gemini 4 Argon vs GLM 5.3
- Gemini 4 Argon vs Gemini 3.7 Flash
- Gemini 4 Argon vs GPT-5.4 Pro
- Gemini 4 Argon vs Claude Opus 4.7
- Gemini 4 Argon vs Claude Opus 4.8
- Gemini 4 Argon vs Qwen3.8 Max
- Gemini 4 Argon vs Gemini 3.5 Flash
- Gemini 4 Argon vs Muse Spark 1.2
- Gemini 4 Argon vs Qwen3.8 2.4T A95B
- Gemini 4 Argon vs DeepSeek V4.1 Flash 1017
- Gemini 4 Argon vs Muse Spark 1.1
- Gemini 4 Argon vs Claude Opus 4.6
- Gemini 4 Argon vs GLM 5.3 Flash
- Gemini 4 Argon vs DeepSeek V4 Pro
- Gemini 4 Argon vs GLM 5.2
- Gemini 4 Argon vs Qwen3.6 Max
- Gemini 4 Argon vs Gemini 3.6 Flash
- Gemini 4 Argon vs MiniMax-M3.1-Flash
- Gemini 4 Argon vs Claude Haiku 5.5
- Gemini 4 Argon vs Muse Spark Open
- Gemini 4 Argon vs DeepSeek V4 Flash
- Gemini 4 Argon vs Gemini 3.1 Pro
- Gemini 4 Argon vs Qwen3 7
- Gemini 4 Argon vs GPT-5.2
- Gemini 4 Argon vs Step 5 Preview
- Gemini 4 Argon vs Gemini 3 Pro
- Gemini 4 Argon vs Claude Sonnet 5
- Gemini 4 Argon vs Qwen3.8-Flash-Next
- Gemini 4 Argon vs GPT-6 Luna
- Gemini 4 Argon vs Claude Opus 4.5
- Gemini 4 Argon vs GPT-5.6 Luna
- Gemini 4 Argon vs MiMo-V2.6-Flash
- Gemini 4 Argon vs GPT-5 Pro
- Gemini 4 Argon vs o3
- Gemini 4 Argon vs Kimi K2.6
- Gemini 4 Argon vs GPT-5.2-Codex
See also
Data as of 21 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Gemini 4 Argon: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/gemini-4-argon, data as of 21 Oct 2026.