RankingEffort benchmarks

Gemini 4 Argon · Medium

Compare models

39 measured results · 4 effort settings · 30 benchmark/harness combinations

Model evidence and sources →

Change model

Model

Current configurationSources

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare withModel and configuration

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

AGemini 4 Argon · Medium1 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
Unknown
Hard reasoning
A79.7Supported
Coding
A92.4Preliminary
Human pref
Unknown
Knowledge
Unknown
Multimodal
Unknown
Long context
Unknown

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

Gemini 4 Argon · Medium: 90.2 overall comparison score. Meets the overall evidence requirements.

2 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Shared-result disagreement. This aggregate order conflicts with matched benchmark results against 1 other configurations. Different test coverage and opponents affect the fit. Inspect published results before treating the rank as conclusive.

Which sources carry the weight?

ARC Prize · 70.9% direct comparison weight · independent of the model provider
Cognition · 29.1% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

ARC-AGI · 37.7% direct comparison weight
arc-agi-3 · 33.2% direct comparison weight
FrontierCode · 29.1% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-10-20.

Both ranks use the same surviving configurations for each removal. The cohort starts with 444 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed source	Rank before → after	Rank movement	Cohort remaining / lost support	Score after (change)
Artificial Analysis	7 → 9	Down 2	203 remaining / 241 lost	90.5 (+0.3 points)
ARC Prize	7 → 1	Up 6	444 remaining / 0 lost	97.4 (+7.2 points)
Datacurve	7 → 8	Down 1	444 remaining / 0 lost	88.6 (-1.6 points)
LiveCodeBench	7 → 7	No movement	444 remaining / 0 lost	90.2 (+0.0 points)
Scale	7 → 7	No movement	444 remaining / 0 lost	90.5 (+0.3 points)
Terminal-Bench	7 → 6	Up 1	444 remaining / 0 lost	89.9 (-0.3 points)
Cognition	7 → 9	Down 2	444 remaining / 0 lost	86.4 (-3.8 points)
LiveBench	7 → 8	Down 1	443 remaining / 1 lost	90.6 (+0.5 points)
SWE-rebench	7 → 7	No movement	444 remaining / 0 lost	90.0 (-0.2 points)
Epoch AI	7 → 8	Down 1	431 remaining / 13 lost	90.2 (-0.0 points)
Vals AI	7 → 7	No movement	443 remaining / 1 lost	90.0 (-0.2 points)
Provider reports · Anthropic	7 → 7	No movement	444 remaining / 0 lost	90.2 (-0.0 points)
Provider reports · Moonshot AI	7 → 7	No movement	444 remaining / 0 lost	90.1 (-0.1 points)
Provider reports · Meta	7 → 7	No movement	444 remaining / 0 lost	90.2 (-0.0 points)
Provider reports · Mistral	7 → 7	No movement	444 remaining / 0 lost	90.2 (+0.0 points)
Provider reports · OpenAI	7 → 5	Up 2	444 remaining / 0 lost	90.0 (-0.2 points)
All provider reports	7 → 5	Up 2	444 remaining / 0 lost	89.8 (-0.4 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family

Removed evidence	Recomputed score	Change
arc-agi	94.1	+3.9
arc-agi-3	86.2	-3.9
frontiercode	86.4	-3.8
deepswe	88.7	-1.5
terminal-bench	89.9	-0.2
mmmu-pro	90.4	+0.2
simpleqa	90.4	+0.2
swe-rebench	90.0	-0.2
critpt	90.3	+0.1
hle	90.3	+0.1
ifbench	90.0	-0.1
epoch-game-puzzles	90.0	-0.1
gpqa	90.1	-0.1
frontiermath	90.1	-0.1
swe-atlas-test-writing	90.1	-0.1
mmlu-pro	90.3	+0.1
swe-atlas-qna	90.1	-0.1
tau-bench	90.1	-0.1
multi-swe-bench	90.1	-0.1
automationbench	90.1	-0.1
livebench-instructions	90.3	+0.1
gdp-pdf	90.2	+0.1
gdpval	90.2	+0.1
vibe-code	90.1	-0.0
omniscience	90.2	+0.0
livebench-reasoning	90.2	-0.0
livebench-data	90.2	+0.0
harvey	90.2	-0.0
enigma-eval	90.2	+0.0
aime	90.2	-0.0
vals-excel-modeling	90.2	-0.0
apex-agents	90.2	+0.0
livebench-language	90.2	+0.0
vals-code-migration	90.2	-0.0
swe-atlas-refactoring	90.2	-0.0
briefcase	90.2	+0.0
terminal-bench-science	90.2	-0.0
aa-lcr	90.2	+0.0
livebench-math	90.2	-0.0
livebench-coding	90.2	-0.0
vals-tax-agent	90.2	-0.0
vals-finance-agent	90.2	+0.0
vals-legal-research	90.2	+0.0
livecodebench	90.2	+0.0
itbench	90.2	-0.0
gmmlu	90.2	-0.0
math500	90.2	-0.0
analyst-agent	90.2	-0.0
swe-bench-pro	90.2	-0.0
enterprise-ops	90.2	-0.0

What evidence is missing from the fit?

Across all collected settings for this model: 67 observations, 32 contributing, 0 matched without graph weight and 35 excluded. Counts do not establish rank eligibility.

1 · Benchmark family or source protocol has not been reviewed
1 · Original SciCode grading defects are under review; retained as supporting evidence, excluded from aggregate and capability scores. Corrected SciCode-Verified requires a separate protocol review. Sources: https://arxiv.org/abs/2608.04975 and https://artificialanalysis.ai/evaluations/scicode
4 · Evaluator has not been reviewed for independent admission
1 · Cache hit rate is much lower due to a non-production setting.
18 · Provider-published claim. Independent source results are admitted separately with exact checkpoint, effort and compatible protocol; this grid does not establish a matched comparison.
1 · Provider-published claim. Independent source results are admitted separately with exact checkpoint, effort and compatible protocol; this grid does not establish a matched comparison. Historical provider claim; secondary and superseded for current FrontierSWE comparisons by the independent Proximal FrontierSWE v2 mean@5 board at https://www.frontierswe.com/. Version-unspecified claims are not v2 results. Excluded from weighted Overall.
1 · No reviewed effort for this catalog observation; any separately collected effort measurement is counted separately
7 · Supporting evidence outside the reviewed capability core
1 · No matched opponent in this evaluation unit

Inspect every observation and exclusion →

Benchmark results

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / source	Low	Medium	High	Unspecified
AA-AnalystAgentArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks solved on all 5 of 5 attempts · 80 tasks, 5 attempts per task · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	48.75% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	1,486.44 Elo Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	79.67% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	77.51% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally. CritPt is under review by its authors following external feedback · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	27.14% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	21.8% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-18.	Not reported	Not reported	1,623.99 Elo Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	57.09% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-10-18.	Not reported	Not reported	42.35 Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	49.9% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark curated by scientists across 16 disciplines. SciCode is under review following an independent audit of the dataset · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	61.81% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
Terminal-Bench-Science 0.1Artificial Analysis · mini-swe-agent; Terminal-Bench-Science 0.1.0; 70 tasks; pass@1 averaged over 3 repeats Benchmark details Share of tasks passed (pass@1), averaged over three repeats · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	52.86% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	57.07% Reported settings Gemini 4 Argon (High); Model source: https://artificialanalysis.ai/models/gemini-4-argon	Not reported
ARC-AGI-2ARC Prize · v2 Semi-Private Benchmark details Original public board. Configuration and task coverage may differ; inspect source. Unit: percent. Source collected 2026-10-18.	42.08% Reported settings Gemini 4 Argon (Low); Source model ID: google-gemini-4-argon-low; 120 tasks	87.08% Reported settings Gemini 4 Argon (Medium); Source model ID: google-gemini-4-argon-medium; 120 tasks	91.25% Reported settings Gemini 4 Argon (High); Source model ID: google-gemini-4-argon-high; 120 tasks	Not reported
ARC-AGI-3 · provider-adapterARC Prize · Provider Adapter Benchmark details ARC-AGI-3; Provider Adapter. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	97.46% Reported settings ARC-AGI-3 Semi-Private verified summary; Provider Adapter harness; source group google-gemini-4-argon; variant provider-adapter-low; public-demo task scores are excluded	97.34% Reported settings ARC-AGI-3 Semi-Private verified summary; Provider Adapter harness; source group google-gemini-4-argon; variant provider-adapter-medium; public-demo task scores are excluded	97.3% Reported settings ARC-AGI-3 Semi-Private verified summary; Provider Adapter harness; source group google-gemini-4-argon; variant provider-adapter-high; public-demo task scores are excluded	Not reported
ARC-AGI-3 · standardARC Prize · Standard Benchmark details ARC-AGI-3; Standard. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	49.79% Reported settings ARC-AGI-3 Semi-Private verified summary; Standard harness; source group google-gemini-4-argon; variant low; public-demo task scores are excluded	46.27% Reported settings ARC-AGI-3 Semi-Private verified summary; Standard harness; source group google-gemini-4-argon; variant medium; public-demo task scores are excluded	Not reported	Not reported
FrontierCode 1.1 · extended-chiselCognition · chisel Benchmark details FrontierCode 1.1; chisel. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	60.58% Reported settings FrontierCode 1.1; extended; 150 tasks; chisel; source effort low; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0229	60.5% Reported settings FrontierCode 1.1; extended; 150 tasks; chisel; source effort medium; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0134	59.38% Reported settings FrontierCode 1.1; extended; 150 tasks; chisel; source effort high; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0137	Not reported
FrontierCode 1.1 · main-chiselCognition · chisel Benchmark details FrontierCode 1.1; chisel. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-18.	47.2% Reported settings FrontierCode 1.1; main; 100 tasks; chisel; source effort low; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0304	47.55% Reported settings FrontierCode 1.1; main; 100 tasks; chisel; source effort medium; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0121	45.87% Reported settings FrontierCode 1.1; main; 100 tasks; chisel; source effort high; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate 0.0103	Not reported
Vals AI · Finance Agent v2Vals AI · Vals FAB v2 shared six-tool harness; two-hour task limit Benchmark details Overall weighted rubric accuracy, averaged over three runs on 450 tasks. Dealbreaker failures and timeouts score zero. Related categories and all-pass scores are not additional votes. Three-model judge panel; private test set. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	65.4% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":null}	Not reported
Vals AI · Code Migration v1Vals AI · Vals Code Migration shared agent harness Benchmark details Hidden behavior-test pass rate, equally weighted across 40 repositories (30 CLI and 10 COBOL-to-Java). Anti-cheat failures score zero. Code-quality ratings and language splits are not additional votes. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	68.17% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":null}	Not reported
Vals AI · Legal Research Bench v1Vals AI · Vals Legal Research shared five-tool harness; three-hour task limit Benchmark details All-pass accuracy on 208 private legal research tasks. Every expert rubric check must pass; GPT-5.4 judges responses. Partial-credit scores and practice-area splits are not additional votes. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	54.81% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":null}	Not reported
Vals AI · Excel Modeling Benchmark v1Vals AI · Vals EMB shared workbook agent harness; 3.5-hour task limit Benchmark details Mean of seven equally weighted task categories. Template submissions use recalculated Excel cell accuracy; scratch submissions use rubric checks judged by GPT-5.4 Mini XHigh. Partial credit does not mean task completion. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	75.24% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":null}	Not reported
Vals AI · Vibe Code Bench v1.1Vals AI · Vals modified OpenHands harness; five hours or 1,000 turns per app Benchmark details Mean app test pass rate on the canonical v1.1 test set. Browser Use executes workflow tests; each test requires at least 90% of its substeps. Automated judging can introduce error. Alternate agents are excluded. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	91.91% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":"OpenHands"}	Not reported
Vals AI · Harvey Legal Agent Benchmark v1Vals AI · Vals Harvey held-out legal file-work protocol; internet disabled Benchmark details Mean task resolution across two judges (GPT-5.5 Medium and Sonnet 4.6), requiring every criterion to pass. Criteria pass rates and practice-area splits do not add votes. Shared Harvey family; Fable 5 fallback run excluded. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	19.58% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":null}	Not reported
Vals AI · Tax Agent Bench v1: Weighted PassVals AI · Vals Tax Agent Bench shared five-tool harness; three-hour task limit Benchmark details Weighted Pass on 193 private US corporate-tax questions. Each task score is the weighted rubric share, set to zero when a must-pass check fails, then multiplied by (0.7 + 0.3 times citation quality). Timeouts score zero. Category splits and all-pass rates are not additional votes. GPT-5.4 judges expert rubrics. Unit: percent. Source collected 2026-10-18.	Not reported	Not reported	76.23% Reported settings {"checkpoint":"gemini-4-argon","evaluatorModel":"google/gemini-4-argon","sourceUpdatedOn":"2026-10-16","reasoning_effort":"high","compute_effort":null,"reasoning":null,"temperature":1,"top_p":null,"max_output_tokens":262144,"verbosity":null,"provider":"Google","harness":null}	Not reported

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 8,915 results across 259 models · Collected 2026-10-18

Gemini 4 Argon · Medium

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results