Google DeepMindproprietary
Released: 2026-10-13Gemini 4 Argon
Google DeepMind fourth-generation multimodal flagship with 2M native context and real-time environment grounding.
Intelligence Index
96.0/ 100
Calibrated multi-domain compositeThroughput Speed
90tok/sec
Output streaming throughputTime To First Token
320ms
Initial chunk server latencyLiveBench Contam-Free
84.1%
September 2026 suite scoreSourced Benchmark Evaluations (61)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| RULER | Nearing Saturation | 88.3% | 84 | 2026-09-12 | independent | Source ↗ |
| Mercury | Deprecated | 78.3% | 10 | 2026-09-12 | independent | Source ↗ |
| MGSM | Nearing Saturation | 90.3% | 65 | 2026-09-12 | independent | Source ↗ |
| MMLU | Saturated | 97.9% | 25 | 2026-09-12 | independent | Source ↗ |
| MMLU-Pro | Nearing Saturation | 88.3% | 75 | 2026-09-12 | independent | Source ↗ |
| MMLU-Redux | Saturated | 94.1% | 25 | 2026-09-12 | independent | Source ↗ |
| MT-Bench | Nearing Saturation | 9.1% | 75 | 2026-09-12 | independent | Source ↗ |
| MultiPL-E | Deprecated | 81.9% | 10 | 2026-09-12 | independent | Source ↗ |
| Needle-in-a-Haystack | Saturated | 99.2% | 44 | 2026-09-12 | independent | Source ↗ |
| OlympiadBench | Active | 77.9% | 80 | 2026-09-12 | independent | Source ↗ |
| Omni-MATH | Deprecated | 70.3% | 10 | 2026-09-12 | independent | Source ↗ |
| OpenAI-MRCR | Active | 80.6% | 80 | 2026-09-12 | independent | Source ↗ |
| PutnamBench | Active | 49.4% | 90 | 2026-09-12 | independent | Source ↗ |
| QASPER | Nearing Saturation | 79.7% | 65 | 2026-09-12 | independent | Source ↗ |
| QuALITY | Nearing Saturation | 84.5% | 78 | 2026-09-12 | independent | Source ↗ |
| AGIEval | Saturated | 82.4% | 25 | 2026-09-12 | independent | Source ↗ |
| SCROLLS | Nearing Saturation | 76.8% | 65 | 2026-09-12 | independent | Source ↗ |
| SimpleQA | Active | 49.9% | 90 | 2026-09-12 | independent | Source ↗ |
| Soohak | Active | 72.2% | 98 | 2026-09-12 | independent | Source ↗ |
| StructFlowBench | Nearing Saturation | 75.1% | 75 | 2026-09-12 | independent | Source ↗ |
| SuperGLUE | Saturated | 97.7% | 25 | 2026-09-12 | independent | Source ↗ |
| SWE-bench | Deprecated | 60.5% | 10 | 2026-09-12 | independent | Source ↗ |
| SWE-bench Pro | Active | 55.2% | 90 | 2026-09-12 | independent | Source ↗ |
| SWE-bench Verified | Saturated | 69.4% | 25 | 2026-09-12 | independent | Source ↗ |
| SWE-Lancer | Deprecated | 60.5% | 20 | 2026-09-12 | independent | Source ↗ |
| TabMWP | Saturated | 87.4% | 25 | 2026-09-12 | independent | Source ↗ |
| Terminal-Bench | Active | 62.3% | 80 | 2026-09-12 | independent | Source ↗ |
| TruthfulQA | Saturated | 84.5% | 25 | 2026-09-12 | independent | Source ↗ |
| WinoGrande | Saturated | 99.6% | 28 | 2026-09-12 | independent | Source ↗ |
| ZeroSCROLLS | Nearing Saturation | 76.8% | 78 | 2026-09-12 | independent | Source ↗ |
| MBPP | Saturated | 94.3% | 25 | 2026-09-12 | independent | Source ↗ |
| AgentIF | Nearing Saturation | 60.1% | 80 | 2026-09-12 | independent | Source ↗ |
| AIME | Nearing Saturation | 90.3% | 75 | 2026-09-12 | independent | Source ↗ |
| ARC-AGI | Saturated | 90.1% | 25 | 2026-09-12 | independent | Source ↗ |
| ARC-AGI-2 | Nearing Saturation | 74.7% | 75 | 2026-09-12 | independent | Source ↗ |
| ARC-AGI-3 | Active | 62.3% | 93 | 2026-09-12 | independent | Source ↗ |
| BIG-Bench Hard | Saturated | 93.9% | 25 | 2026-09-12 | independent | Source ↗ |
| CodeContests | Deprecated | 57.9% | 10 | 2026-09-12 | independent | Source ↗ |
| CodeXGLUE | Deprecated | 83.7% | 23 | 2026-09-12 | independent | Source ↗ |
| CommonsenseQA | Saturated | 100% | 28 | 2026-09-12 | independent | Source ↗ |
| ComplexBench | Nearing Saturation | 72.4% | 80 | 2026-09-12 | independent | Source ↗ |
| CyberSecEval | Deprecated | 80.1% | 10 | 2026-09-12 | independent | Source ↗ |
| FireBench | Active | 74.2% | 95 | 2026-09-12 | independent | Source ↗ |
| FrontierCode | Active | 51.6% | 98 | 2026-09-12 | independent | Source ↗ |
| FrontierMath | Active | 26.6% | 98 | 2026-09-12 | independent | Source ↗ |
| GLUE | Saturated | 100% | 28 | 2026-09-12 | independent | Source ↗ |
| GPQA | Nearing Saturation | 84.5% | 78 | 2026-09-12 | independent | Source ↗ |
| GSM-Hard | Deprecated | 91.2% | 10 | 2026-09-12 | independent | Source ↗ |
| GSM8K | Saturated | 100% | 25 | 2026-09-12 | independent | Source ↗ |
| HumanEval | Saturated | 100% | 25 | 2026-09-12 | independent | Source ↗ |
| HumanEval+ | Saturated | 96.1% | 25 | 2026-09-12 | independent | Source ↗ |
| Humanity's Last Exam | Active | 42.2% | 93 | 2026-09-12 | independent | Source ↗ |
| IFEval | Deprecated | 86.6% | 25 | 2026-09-12 | independent | Source ↗ |
| InfiniteBench | Nearing Saturation | 81.6% | 65 | 2026-09-12 | independent | Source ↗ |
| LiveCodeBench | Saturated | 73% | 39 | 2026-09-12 | independent | Source ↗ |
| LongBench | Nearing Saturation | 82.6% | 65 | 2026-09-12 | independent | Source ↗ |
| LongBench-Pro | Active | 78.7% | 80 | 2026-09-12 | independent | Source ↗ |
| MATH | Saturated | 93.1% | 25 | 2026-09-12 | independent | Source ↗ |
| MathArena | Active | 76% | 99 | 2026-09-12 | independent | Source ↗ |
| MBPP+ | Nearing Saturation | 90.8% | 65 | 2026-09-12 | independent | Source ↗ |
| LiveBench (Continuous) | Active | 84.1% | 94 | 2025-02-01 | independent | Source ↗ |
OpenAI Compatible API ExecutionAPI Ready
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gemini-4-argon",
messages=[{"role": "user", "content": "Evaluate multi-step logic problem"}],
temperature=0.2
)
print(response.choices[0].message.content)Evaluation & Sourcing Notes
Scores listed for Gemini 4 Argon represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.