Head-to-head comparison
GPT-5.4 vs. Grok 4.6
Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.
53
shared result series
16
distinct benchmarks
2026-08-16
latest retrieval
Change model selection
Specifications and cost
General specifications and standard API pricing
Undisclosed parameter counts and knowledge cutoffs stay marked as unavailable. The table shows direct standard token pricing, including documented context tiers. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.
| Attribute | GPT-5.4 | Grok 4.6 |
|---|---|---|
| Provider | OpenAI | xAI |
| Released | Mar 2026 | Aug 12, 2026 |
| Availability | Active | Active |
| Model type | Proprietary | Proprietary |
| Parameters | Not published | Not published |
| Architecture | Not published | Not published |
| Context window | 1,050,000 tokens | 500,000 tokens |
| Knowledge cutoff | August 31, 2025 | February 1, 2026 |
| Technical sources | Model | Model |
| API pricing per 1M tokens | ||
| API input | $2.5 ≤272K / $5 >272K | $2 ≤200K / $4 >200K |
| API output | $15 ≤272K / $22.5 >272K | $6 ≤200K / $12 >200K |
| Cache read | $0.25 ≤272K / $0.5 >272K | $0.5 ≤200K / $1 >200K |
| Cache write (5 min.) | Not listed | Not listed |
| Cache write (1 hr.) | Not listed | Not listed |
| Price verified | 07/30/2026Source | 08/16/2026Source |
GPT-5.4Grok 4.6
API inputUSD per 1M tokens, base tier
GPT-5.4$2.5
Grok 4.6$2Lower price
API outputUSD per 1M tokens, base tier
GPT-5.4$15
Grok 4.6$6Lower price
Cache readUSD per 1M tokens, base tier
GPT-5.4$0.25Lower price
Grok 4.6$0.5
Performance
Shared benchmarks
Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.
53 shared result series
GPT-5.4Grok 4.6
Arena Text, Style ControlOverall. Higher is better. Test setup not documented as identical
GPT-5.41,465.439 arena pointsHigher measured value
Grok 4.61,464.323 arena points
Code MigrationOverall. Higher is better. No documented setup difference
GPT-5.434.984 %
Grok 4.644.572 %Winner
CorpFin v2Overall. Higher is better. No documented setup difference
GPT-5.465.268 %
Grok 4.666.161 %Winner
GPQA DiamondOverall. Higher is better. No documented setup difference
GPT-5.491.666 %
Grok 4.694.697 %Winner
Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. No documented setup difference
GPT-5.40 %
Grok 4.615.833 %Winner
LegalBenchOverall. Higher is better. No documented setup difference
GPT-5.486.044 %
Grok 4.686.307 %Winner
LiveCodeBench v6Overall. Higher is better. No documented setup difference
GPT-5.484.141 %
Grok 4.688.224 %Winner
MedCodeOverall. Higher is better. No documented setup difference
GPT-5.441.292 %
Grok 4.644.712 %Winner
MedScribeOverall. Higher is better. No documented setup difference
GPT-5.477.549 %
Grok 4.686.534 %Winner
MMLU-ProOverall. Higher is better. No documented setup difference
GPT-5.487.482 %
Grok 4.689.4 %Winner
MortgageTaxOverall. Higher is better. No documented setup difference
GPT-5.468.323 %Winner
Grok 4.664.189 %
SAGEOverall. Higher is better. No documented setup difference
GPT-5.443.312 %Winner
Grok 4.628.9 %
| Benchmark and task | GPT-5.4 | Grok 4.6 | Comparability | Source |
|---|---|---|---|---|
| Arena Text, Style ControlOverall | 1,465.439 arena pointsHigher measured value 95% confidence interval: ±3.814The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,464.323 arena points 95% confidence interval: ±11.88The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:overallTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 63,759 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 2,448 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Code MigrationOverall | 34.984 % | 44.572 %Winner | 1accuracyHigher is betterCohort: vals-code-migration:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-13Import ID: vals-code-migration-2026-08-16-107652dc6bfcRetrieved 2026-08-16 |
| CorpFin v2Overall | 65.268 % | 66.161 %Winner | 2accuracyHigher is betterCohort: vals-corp-fin-v2:2:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-corp_fin_v2-2026-08-16-45bbf84f57abRetrieved 2026-08-16 |
| GPQA DiamondOverall | 91.666 % | 94.697 %Winner | 1accuracyHigher is betterCohort: vals-gpqa:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-gpqa-2026-08-16-7f8363eaa2dbRetrieved 2026-08-16 |
| Harvey's Legal Agent BenchmarkOverall · Task fully resolved | 0 % | 15.833 %Winner | 1task resolution rateHigher is betterCohort: vals-legal-agent-benchmark:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-14Import ID: vals-hlab-2026-08-16-bae90566df2fRetrieved 2026-08-16 |
| LegalBenchOverall | 86.044 % | 86.307 %Winner | 1accuracyHigher is betterCohort: vals-legal-bench:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-legal_bench-2026-08-16-d5b3b1b2b5c8Retrieved 2026-08-16 |
| LiveCodeBench v6Overall | 84.141 % | 88.224 %Winner | 1accuracyHigher is betterCohort: vals-lcb:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-13Import ID: vals-lcb-2026-08-16-17b9c147080dRetrieved 2026-08-16 |
| MedCodeOverall | 41.292 % | 44.712 %Winner | 1accuracyHigher is betterCohort: vals-medcode:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-medcode-2026-08-16-524f7f474eacRetrieved 2026-08-16 |
| MedScribeOverall | 77.549 % | 86.534 %Winner | 1accuracyHigher is betterCohort: vals-medscribe:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-medscribe-2026-08-16-c385354a7e62Retrieved 2026-08-16 |
| MMLU-ProOverall | 87.482 % | 89.4 %Winner | 1accuracyHigher is betterCohort: vals-mmlu-pro:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-mmlu_pro-2026-08-16-e14d5bb3a34cRetrieved 2026-08-16 |
| MortgageTaxOverall | 68.323 %Winner | 64.189 % | 1accuracyHigher is betterCohort: vals-mortgage-tax:1:overallNo documented setup differenceWinner: GPT-5.4GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-mortgage_tax-2026-08-16-e78c985dbb1cRetrieved 2026-08-16 |
| SAGEOverall | 43.312 %Winner | 28.9 % | 1accuracyHigher is betterCohort: vals-sage:1:overallNo documented setup differenceWinner: GPT-5.4GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-sage-2026-08-16-a416dd25affbRetrieved 2026-08-16 |
| SkillsBenchOverall | 51.711 % | 55.766 %Winner | 1accuracyHigher is betterCohort: vals-skillsbench:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05, Harness: OpenHands Grok 4.6: Harness: OpenHands | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-13Import ID: vals-skillsbench-2026-08-16-4917a67268d2Retrieved 2026-08-16 |
| SWE-bench VerifiedOverall | 78.2 % | 95.6 %Winner | 1accuracyHigher is betterCohort: vals-swebench:1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-swebench-2026-08-16-c133071b8167Retrieved 2026-08-16 |
| TaxEval v2Overall | 73.958 %Winner | 71.096 % | 2accuracyHigher is betterCohort: vals-tax-eval-v2:2:overallNo documented setup differenceWinner: GPT-5.4GPT-5.4: Snapshot: gpt-5.4-2026-03-05 Grok 4.6: No further settings disclosed | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-12Import ID: vals-tax_eval_v2-2026-08-16-5bbd495dac53Retrieved 2026-08-16 |
| Vibe Code Bench v1.1Overall | 67.421 % | 76.239 %Winner | 1.1accuracyHigher is betterCohort: vals-vibe-code-bench:1.1:overallNo documented setup differenceWinner: Grok 4.6GPT-5.4: Snapshot: gpt-5.4-2026-03-05, Harness: OpenHands Grok 4.6: Harness: OpenHands | Vals AIType: Independent evaluationLeaderboard updated: 2026-08-13Import ID: vals-vibe-code-2026-08-16-a73c9770bcc4Retrieved 2026-08-16 |
| Arena Text, Style ControlOverallReasoning: high | 1,476.571 arena pointsNo clear lead 95% confidence interval: ±3.887The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,464.323 arena points 95% confidence interval: ±11.88The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:overallNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 60,819 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 2,448 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlBusiness, management and finance | 1,474.218 arena pointsHigher measured value 95% confidence interval: ±6.56The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,455.102 arena points 95% confidence interval: ±27.332The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_business_and_management_and_financial_operationsTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 13,072 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 442 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlBusiness, management and financeReasoning: high | 1,483.617 arena pointsNo clear lead 95% confidence interval: ±6.683The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,455.102 arena points 95% confidence interval: ±27.332The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_business_and_management_and_financial_operationsNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 12,373 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 442 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlChinese | 1,496.688 arena points 95% confidence interval: ±11.202The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,535.167 arena pointsHigher measured value 95% confidence interval: ±39.462The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:chineseTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 3,574 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 215 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlChineseReasoning: high | 1,506.007 arena points 95% confidence interval: ±11.398The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,535.167 arena pointsNo clear lead 95% confidence interval: ±39.462The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:chineseNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 3,369 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 215 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlCoding | 1,513.75 arena pointsHigher measured value 95% confidence interval: ±5.977The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,511.77 arena points 95% confidence interval: ±26.72The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:codingTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 17,752 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 485 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlCodingReasoning: high | 1,520.704 arena pointsNo clear lead 95% confidence interval: ±6.02The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,511.77 arena points 95% confidence interval: ±26.72The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:codingNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 16,515 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 485 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlCreative writing | 1,436.291 arena points 95% confidence interval: ±7.077The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,472.894 arena pointsHigher measured value 95% confidence interval: ±31.578The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:creative_writingTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 10,314 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 379 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlCreative writingReasoning: high | 1,444.387 arena points 95% confidence interval: ±7.221The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,472.894 arena pointsNo clear lead 95% confidence interval: ±31.578The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:creative_writingNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 10,119 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 379 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlEnglish | 1,468.758 arena pointsHigher measured value 95% confidence interval: ±4.955The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,459.266 arena points 95% confidence interval: ±18.461The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:englishTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 29,658 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 998 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlEnglishReasoning: high | 1,478.413 arena pointsNo clear lead 95% confidence interval: ±5.006The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,459.266 arena points 95% confidence interval: ±18.461The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:englishNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 28,120 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 998 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlEntertainment, sports and media | 1,433.639 arena points 95% confidence interval: ±6.465The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,455.476 arena pointsHigher measured value 95% confidence interval: ±28.81The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_entertainment_and_sports_and_mediaTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 13,629 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 448 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlEntertainment, sports and mediaReasoning: high | 1,443.099 arena points 95% confidence interval: ±6.632The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,455.476 arena pointsNo clear lead 95% confidence interval: ±28.81The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_entertainment_and_sports_and_mediaNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 12,966 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 448 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlExcluding ties | 1,471.723 arena pointsHigher measured value 95% confidence interval: ±4.974The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,469.633 arena points 95% confidence interval: ±16.492The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:exclude_tiesTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 48,930 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,782 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlExcluding tiesReasoning: high | 1,486.146 arena pointsNo clear lead 95% confidence interval: ±5.068The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,469.633 arena points 95% confidence interval: ±16.492The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:exclude_tiesNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 46,537 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,782 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlExpert prompts | 1,490.915 arena pointsHigher measured value 95% confidence interval: ±8.635The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,479.969 arena points 95% confidence interval: ±34.191The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:expertTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 6,142 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 272 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlExpert promptsReasoning: high | 1,516.663 arena pointsNo clear lead 95% confidence interval: ±8.789The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,479.969 arena points 95% confidence interval: ±34.191The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:expertNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 5,794 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 272 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlHard prompts | 1,487.682 arena points 95% confidence interval: ±4.603The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,488.551 arena pointsHigher measured value 95% confidence interval: ±15.88The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:hard_promptsTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 41,916 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,397 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlHard promptsReasoning: high | 1,498.61 arena pointsNo clear lead 95% confidence interval: ±4.64The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,488.551 arena points 95% confidence interval: ±15.88The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:hard_promptsNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 39,412 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,397 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlHard prompts, English | 1,486.794 arena pointsHigher measured value 95% confidence interval: ±5.668The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,472.181 arena points 95% confidence interval: ±25.031The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:hard_prompts_englishTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 20,340 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 543 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlHard prompts, EnglishReasoning: high | 1,495.669 arena pointsNo clear lead 95% confidence interval: ±5.735The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,472.181 arena points 95% confidence interval: ±25.031The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:hard_prompts_englishNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 18,992 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 543 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlInstruction following | 1,461.6 arena pointsHigher measured value 95% confidence interval: ±5.543The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,460.755 arena points 95% confidence interval: ±21.635The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:instruction_followingTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 21,791 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 732 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlInstruction followingReasoning: high | 1,473.561 arena pointsNo clear lead 95% confidence interval: ±5.61The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,460.755 arena points 95% confidence interval: ±21.635The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:instruction_followingNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 20,349 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 732 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlLife, physical and social science | 1,478.92 arena points 95% confidence interval: ±6.927The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,497.908 arena pointsHigher measured value 95% confidence interval: ±31.96The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_life_and_physical_and_social_scienceTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 10,340 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 335 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlLife, physical and social scienceReasoning: high | 1,488.258 arena points 95% confidence interval: ±7.051The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,497.908 arena pointsNo clear lead 95% confidence interval: ±31.96The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_life_and_physical_and_social_scienceNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 9,911 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 335 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlLonger queries | 1,478.227 arena points 95% confidence interval: ±5.383The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,485.451 arena pointsHigher measured value 95% confidence interval: ±18.11The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:longer_queryTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 27,468 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,103 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlLonger queriesReasoning: high | 1,483.605 arena points 95% confidence interval: ±5.458The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,485.451 arena pointsNo clear lead 95% confidence interval: ±18.11The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:longer_queryNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 26,121 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,103 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlMulti-turn | 1,481.026 arena pointsHigher measured value 95% confidence interval: ±6.734The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,444.26 arena points 95% confidence interval: ±31.774The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:multi_turnTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 12,008 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 352 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlMulti-turnReasoning: high | 1,494.427 arena pointsWinner 95% confidence interval: ±6.811The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,444.26 arena points 95% confidence interval: ±31.774The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:multi_turnNo documented setup differenceWinner: GPT-5.4GPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 11,302 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 352 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlNon-English | 1,458.043 arena points 95% confidence interval: ±4.754The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,461.99 arena pointsHigher measured value 95% confidence interval: ±15.246The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:non_englishTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 34,099 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,450 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlNon-EnglishReasoning: high | 1,468.846 arena pointsNo clear lead 95% confidence interval: ±4.8The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,461.99 arena points 95% confidence interval: ±15.246The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:non_englishNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 32,699 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 1,450 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlRussian | 1,471.407 arena pointsHigher measured value 95% confidence interval: ±8.204The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,446.737 arena points 95% confidence interval: ±34.404The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:russianTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 6,797 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 275 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlRussianReasoning: high | 1,490.981 arena pointsWinner 95% confidence interval: ±8.413The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,446.737 arena points 95% confidence interval: ±34.404The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:russianNo documented setup differenceWinner: GPT-5.4GPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 6,401 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 275 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlSoftware and IT services | 1,498.772 arena points 95% confidence interval: ±5.275The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,511.885 arena pointsHigher measured value 95% confidence interval: ±20.78The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_software_and_it_servicesTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 25,369 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 805 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlSoftware and IT servicesReasoning: high | 1,506.788 arena points 95% confidence interval: ±5.317The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,511.885 arena pointsNo clear lead 95% confidence interval: ±20.78The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_software_and_it_servicesNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 23,913 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 805 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlWriting, literature and language | 1,459.832 arena pointsHigher measured value 95% confidence interval: ±6.107The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,455.393 arena points 95% confidence interval: ±25.717The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_writing_and_literature_and_languageTest setup not documented as identicalNo direct winnerGPT-5.4: Snapshot: gpt-5.4, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 15,400 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 543 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
| Arena Text, Style ControlWriting, literature and languageReasoning: high | 1,466.108 arena pointsNo clear lead 95% confidence interval: ±6.163The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | 1,455.393 arena points 95% confidence interval: ±25.717The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result. | style-control-v1style-controlled Bradley-Terry ratingHigher is betterCohort: lmarena-text-style-control:2026-08-12:industry_writing_and_literature_and_languageNo documented setup differenceNo clear leadGPT-5.4: Snapshot: gpt-5.4-high, Reasoning: high, Harness: Arena text, style control Grok 4.6: Snapshot: grok-4.6-high, Reasoning: high, Harness: Arena text, style controlGPT-5.4: Style-controlled Arena rating from 15,044 anonymous pairwise votes.Grok 4.6: Style-controlled Arena rating from 543 anonymous pairwise votes. | ArenaType: Official benchmark leaderboardLeaderboard updated: 2026-08-12Import ID: lmarena-eeedf9209373a7f9b324bfb839e7b02fedacd1e5-1611ed64f3f1Retrieved 2026-08-16 |
How to read this comparison
Winning one benchmark is not an overall verdict
Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.
Prices use official API rates per 1M tokens. Web subscriptions and cloud platform surcharges can differ.
Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.
More matchups
Related LLM comparisons
Compare GPT-5.4 and Grok 4.6 with other leading models using the same data and benchmark logic.
View the complete LLM comparison