Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.
shared result series
62
distinct benchmarks
4
latest retrieval
2026-09-29
Change model selection
Specifications
General specifications
Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.
The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.
Key to the marks:better valuebetter, but not conclusivetie
Pricing by provider
Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.
Provider pricing for GPT-5.5 and DeepSeek-V4-Flash
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.
Capability profile from matched benchmarks
Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.
GPT-5.5
DeepSeek-V4-Flash
Coding: 100 / 97.8 (1 benchmark family)
Reasoning: 100 / 97.2 (1 benchmark family)
Mathematics: 100 / 95.8 (1 benchmark family)
Knowledge: 100 / 97.4 (1 benchmark family)
Finance: 100 / 96.6 (1 benchmark family)
Legal: 100 / 97.5 (1 benchmark family)
Cybersecurity: 100 / 82.8 (1 benchmark family)
Health: 100 / 98.6 (1 benchmark family)
60 matched rows from 2 benchmark families were considered. The radar shows 8 of 9 comparable categories.
GPT-5.5DeepSeek-V4-Flash
AlmanBench v0.1Accepted cases. Higher is better. Test setup not documented as identical
GPT-5.5Higher measured value95.432 %
DeepSeek-V4-Flash93.003 %
CyberBench v1.1Patch. Higher is better. No documented setup difference
GPT-5.5Winner81.356 %
DeepSeek-V4-Flash72.414 %
LMArena Text, Style ControlOverall. Higher is better. No documented setup difference
GPT-5.5Winner1,476.671
DeepSeek-V4-Flash1,436.082
Taste-BenchOverall. Higher is better. Test setup not documented as identical
GPT-5.5Higher measured value59.5 %
DeepSeek-V4-Flash43.3 %
CyberBench v1.1PoC. Higher is better. No documented setup difference
GPT-5.5Winner79.661 %
DeepSeek-V4-Flash61.017 %
LMArena Text, Style ControlOverall. Higher is better. No documented setup difference
GPT-5.5Winner1,481.397
DeepSeek-V4-Flash1,438.511
LMArena Text, Style ControlBusiness, management and finance. Higher is better. No documented setup difference
GPT-5.5Winner1,485.504
DeepSeek-V4-Flash1,438.526
LMArena Text, Style ControlBusiness, management and finance. Higher is better. No documented setup difference
GPT-5.5Winner1,489.4
DeepSeek-V4-Flash1,434.189
LMArena Text, Style ControlChinese. Higher is better. No documented setup difference
GPT-5.5Winner1,526.973
DeepSeek-V4-Flash1,471.267
LMArena Text, Style ControlChinese. Higher is better. No documented setup difference
GPT-5.5Winner1,517.408
DeepSeek-V4-Flash1,479.924
LMArena Text, Style ControlCoding. Higher is better. No documented setup difference
GPT-5.5Winner1,510.389
DeepSeek-V4-Flash1,484.094
LMArena Text, Style ControlCoding. Higher is better. No documented setup difference
GPT-5.5Winner1,518.674
DeepSeek-V4-Flash1,479.555
Key to the marks:better valuebetter, but not conclusivetie
Benchmark scores for GPT-5.5 and DeepSeek-V4-Flash
Benchmark and task
GPT-5.5
DeepSeek-V4-Flash
Source
AlmanBench v0.1Accepted casesTest details
Test setup not documented as identicalNo direct winnerVersion: v0.1, results commit f34ddbfMetric: accepted casesScoring: Higher is betterGPT-5.5: Reasoning: xhigh DeepSeek-V4-Flash: Reasoning: max1,029 public tasks with one direct run per row. The target is the Alman language specification.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 70,635 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 51,871 anonymous pairwise votes.
Winner1,476.671
95% confidence interval: ±3.701The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,436.082
95% confidence interval: ±4.048The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
Test setup not documented as identicalNo direct winnerVersion: arxiv:2609.25804v2Metric: both-orders-correct, engineering/research macro averageScoring: Higher is betterGPT-5.5: Reasoning: xhigh, Output limit: 65,536 tokens, Attempts: 502 questions, two option orders each DeepSeek-V4-Flash: Reasoning: max, Output limit: 65,536 tokens, Attempts: 502 questions, two option orders eachSame prompt for all models. Each of 502 questions is shown twice with the options reversed. Average of engineering and research success rates, requiring both orders to be correct. 65,536-token output budget; unparseable answers count as wrong.
Higher measured value59.5 %
0 of 1,004 responses were unparseable and counted as wrong. Not a Gradually test.
43.3 %
2 of 1,004 responses were unparseable and counted as wrong. Not a Gradually test.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 69,008 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 51,747 anonymous pairwise votes.
Winner1,481.397
95% confidence interval: ±3.734The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,438.511
95% confidence interval: ±4.115The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlBusiness, management and financeTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 14,097 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 10,436 anonymous pairwise votes.
Winner1,485.504
95% confidence interval: ±6.33The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,438.526
95% confidence interval: ±6.971The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlBusiness, management and financeTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 13,739 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 10,641 anonymous pairwise votes.
Winner1,489.4
95% confidence interval: ±6.348The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,434.189
95% confidence interval: ±6.982The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 4,480 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,996 anonymous pairwise votes.
Winner1,526.973
95% confidence interval: ±9.82The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,471.267
95% confidence interval: ±11.424The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 4,507 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,992 anonymous pairwise votes.
Winner1,517.408
95% confidence interval: ±9.644The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,479.924
95% confidence interval: ±11.42The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 20,022 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 15,262 anonymous pairwise votes.
Winner1,510.389
95% confidence interval: ±5.641The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,484.094
95% confidence interval: ±6.23The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 19,166 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 15,239 anonymous pairwise votes.
Winner1,518.674
95% confidence interval: ±5.731The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,479.555
95% confidence interval: ±6.265The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 13,078 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 8,539 anonymous pairwise votes.
Winner1,448.977
95% confidence interval: ±6.62The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,407.791
95% confidence interval: ±7.543The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 12,802 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 8,638 anonymous pairwise votes.
Winner1,451.747
95% confidence interval: ±6.64The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,410.703
95% confidence interval: ±7.575The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 31,818 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 23,008 anonymous pairwise votes.
Winner1,482.617
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.
1,444.92
95% confidence interval: ±5.29The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 31,089 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 22,908 anonymous pairwise votes.
Winner1,484.099
95% confidence interval: ±4.815The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,449.94
95% confidence interval: ±5.323The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlEntertainment, sports and mediaTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 16,859 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 11,179 anonymous pairwise votes.
Winner1,447.733
95% confidence interval: ±6.067The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,404.567
95% confidence interval: ±6.937The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlEntertainment, sports and mediaTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 16,373 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 11,428 anonymous pairwise votes.
Winner1,448.374
95% confidence interval: ±6.114The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,408.063
95% confidence interval: ±6.896The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 53,872 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 39,226 anonymous pairwise votes.
Winner1,486.08
95% confidence interval: ±4.841The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,431.809
95% confidence interval: ±5.308The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 52,376 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 39,391 anonymous pairwise votes.
Winner1,492.092
95% confidence interval: ±4.849The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,436.008
95% confidence interval: ±5.353The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 7,644 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 5,392 anonymous pairwise votes.
Winner1,511.193
95% confidence interval: ±7.816The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,457.264
95% confidence interval: ±8.778The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 7,384 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 5,311 anonymous pairwise votes.
Winner1,512.621
95% confidence interval: ±7.838The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,466.306
95% confidence interval: ±8.867The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 2,515 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,006 anonymous pairwise votes.
Winner1,501.116
95% confidence interval: ±14.009The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,448.483
95% confidence interval: ±15.039The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 2,522 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,966 anonymous pairwise votes.
Winner1,496.18
95% confidence interval: ±13.947The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,459.263
95% confidence interval: ±15.286The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,104 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 843 anonymous pairwise votes.
Winner1,487.952
95% confidence interval: ±19.226The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,424.769
95% confidence interval: ±21.221The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,052 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 906 anonymous pairwise votes.
No clear lead1,470.538
95% confidence interval: ±19.232The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,433.595
95% confidence interval: ±20.539The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 47,240 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 34,753 anonymous pairwise votes.
Winner1,499.213
95% confidence interval: ±4.37The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,458.919
95% confidence interval: ±4.767The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 45,517 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 34,700 anonymous pairwise votes.
Winner1,501.268
95% confidence interval: ±4.397The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,458.083
95% confidence interval: ±4.832The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 21,852 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 16,176 anonymous pairwise votes.
Winner1,500.03
95% confidence interval: ±5.513The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,463.433
95% confidence interval: ±6.065The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 21,097 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 16,039 anonymous pairwise votes.
Winner1,499
95% confidence interval: ±5.544The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,464.61
95% confidence interval: ±6.133The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 24,807 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 17,961 anonymous pairwise votes.
Winner1,475.139
95% confidence interval: ±5.271The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,428.187
95% confidence interval: ±5.811The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 24,017 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 18,051 anonymous pairwise votes.
Winner1,477.914
95% confidence interval: ±5.323The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,433.413
95% confidence interval: ±5.881The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 799 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 647 anonymous pairwise votes.
Winner1,478.913
95% confidence interval: ±23.332The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,398.363
95% confidence interval: ±24.906The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 899 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 635 anonymous pairwise votes.
Winner1,504.533
95% confidence interval: ±21.865The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,424.242
95% confidence interval: ±25.166The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,178 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 898 anonymous pairwise votes.
No clear lead1,430.608
95% confidence interval: ±18.898The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,392.925
95% confidence interval: ±20.977The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,184 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 951 anonymous pairwise votes.
Winner1,445.57
95% confidence interval: ±18.604The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,378.153
95% confidence interval: ±20.872The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlLegal and governmentTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 5,923 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 4,139 anonymous pairwise votes.
Winner1,485.605
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.
1,447.478
95% confidence interval: ±9.874The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlLegal and governmentTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 5,605 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 4,107 anonymous pairwise votes.
Winner1,494.595
95% confidence interval: ±8.853The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,457.491
95% confidence interval: ±9.927The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlLife, physical and social scienceTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 11,817 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 8,423 anonymous pairwise votes.
Winner1,496.608
95% confidence interval: ±6.545The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,456.345
95% confidence interval: ±7.388The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlLife, physical and social scienceTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 11,621 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 8,509 anonymous pairwise votes.
Winner1,498.219
95% confidence interval: ±6.591The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,459.724
95% confidence interval: ±7.418The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 33,297 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 23,359 anonymous pairwise votes.
Winner1,486.565
95% confidence interval: ±5.004The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,447.83
95% confidence interval: ±5.578The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 31,935 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 23,501 anonymous pairwise votes.
Winner1,487.567
95% confidence interval: ±5.058The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,448.636
95% confidence interval: ±5.577The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 3,498 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,688 anonymous pairwise votes.
Winner1,499.959
95% confidence interval: ±10.564The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,425.467
95% confidence interval: ±11.801The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 3,474 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,619 anonymous pairwise votes.
Winner1,493.341
95% confidence interval: ±10.42The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,441.412
95% confidence interval: ±11.893The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 4,066 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,975 anonymous pairwise votes.
Winner1,499.198
95% confidence interval: ±10.183The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,434.489
95% confidence interval: ±11.462The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 3,995 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,828 anonymous pairwise votes.
Winner1,506.624
95% confidence interval: ±10.147The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,445.003
95% confidence interval: ±11.761The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlMedicine and healthcareTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 5,331 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 3,699 anonymous pairwise votes.
Winner1,486.508
95% confidence interval: ±9.127The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,459.257
95% confidence interval: ±10.503The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlMedicine and healthcareTest details
No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 5,219 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 3,774 anonymous pairwise votes.
No clear lead1,476.143
95% confidence interval: ±9.066The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,461.307
95% confidence interval: ±10.453The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 12,377 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 9,363 anonymous pairwise votes.
Winner1,483.927
95% confidence interval: ±6.623The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,453.414
95% confidence interval: ±7.284The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 11,834 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 9,233 anonymous pairwise votes.
Winner1,487.824
95% confidence interval: ±6.74The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,446.332
95% confidence interval: ±7.34The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 38,816 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 28,862 anonymous pairwise votes.
Winner1,466.059
95% confidence interval: ±4.493The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,423.239
95% confidence interval: ±4.947The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 37,918 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 28,837 anonymous pairwise votes.
Winner1,473.724
95% confidence interval: ±4.54The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,423.125
95% confidence interval: ±5.023The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,199 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,030 anonymous pairwise votes.
No clear lead1,471.041
95% confidence interval: ±17.669The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,438.601
95% confidence interval: ±18.712The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,168 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,037 anonymous pairwise votes.
Winner1,479.177
95% confidence interval: ±17.666The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,434.977
95% confidence interval: ±18.457The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 7,408 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 5,750 anonymous pairwise votes.
Winner1,478.651
95% confidence interval: ±7.718The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,433.519
95% confidence interval: ±8.502The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 7,356 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 5,790 anonymous pairwise votes.
Winner1,481.011
95% confidence interval: ±7.716The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,437.327
95% confidence interval: ±8.593The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlSoftware and IT servicesTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 28,314 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 21,360 anonymous pairwise votes.
Winner1,505.747
95% confidence interval: ±5.024The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,470.864
95% confidence interval: ±5.508The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlSoftware and IT servicesTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 27,465 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 21,329 anonymous pairwise votes.
Winner1,510.207
95% confidence interval: ±5.045The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,472.722
95% confidence interval: ±5.539The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 2,075 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,596 anonymous pairwise votes.
No clear lead1,460.308
95% confidence interval: ±14.321The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,433.099
95% confidence interval: ±16.059The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 1,974 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,497 anonymous pairwise votes.
Winner1,470.641
95% confidence interval: ±14.515The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,429.802
95% confidence interval: ±16.45The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlWriting, literature and languageTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 18,231 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 12,557 anonymous pairwise votes.
Winner1,466.605
95% confidence interval: ±5.845The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,415.672
95% confidence interval: ±6.557The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
LMArena Text, Style ControlWriting, literature and languageTest details
No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterGPT-5.5: Snapshot: gpt-5.5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlGPT-5.5: Style-controlled Arena rating from 17,866 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 12,763 anonymous pairwise votes.
Winner1,469.628
95% confidence interval: ±5.86The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
1,422.277
95% confidence interval: ±6.549The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.
Key to the marks:better valuebetter, but not conclusivetie
Our model tests
Our tests
Compare how the models respond to the same prompt. Each test shows the first attempt, with no subsequent fixes to the generated code. These results do not contribute to an overall score.
Pelican on a bicycle
A community classic for free-form SVG drawing.
GPT-5.5
The API response was not fully received. The model was not evaluated.
Reasoning (requested): high
First attempt
Run date
7 Sept 2026
API provider
openrouter
Requested API model
openai/gpt-5.5
Reasoning setting
high
Pinned endpoint
OpenAI | openai/gpt-5.5-20260423
Temperature
Provider default
Token limit including reasoning
8,192 tokens
Test protocol
community-visual-tests-v1
Technical checks
Failed
Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.
DeepSeek-V4-Flash
API access was unavailable for this run. The model was not evaluated.
Reasoning (requested): high
First attempt
Run date
7 Sept 2026
API provider
openrouter
Requested API model
deepseek/deepseek-v4-flash-0731
Reasoning setting
high
Pinned endpoint
DeepSeek | deepseek/deepseek-v4-flash-20260731
Temperature
Provider default
Token limit including reasoning
8,192 tokens
Test protocol
community-visual-tests-v1
Technical checks
Not run
Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.
Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.
The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.
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-5.5 and DeepSeek-V4-Flash with other leading models using the same data and benchmark logic.