Head-to-head comparison

Claude Opus 5 vs. DeepSeek-V4-Flash

Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.

shared result series: 33
distinct benchmarks: 3
latest retrieval: 2026-09-04

First modelSecond model

Specifications

General specifications

Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.

Specifications for Claude Opus 5 and DeepSeek-V4-Flash
Attribute	Claude Opus 5	DeepSeek-V4-Flash
Provider	Anthropic	DeepSeek
Released	Jul 24, 2026	Apr 24, 2026
Availability	Active	Preview
Model type	Proprietary	Open Weights
Parameters	Not published	284B, 13B active
Architecture	Not published	Mixture of Experts
Context window	1,000,000 tokens	1,000,000 tokens
Knowledge cutoff	May 2026	Not published
Technical sources	Model, Knowledge cutoff	Model

Pricing

API pricing compared

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.

API pricing for Claude Opus 5 and DeepSeek-V4-Flash
Price type	Claude Opus 5	DeepSeek-V4-Flash
API input	$5	$0.14
API output	$25	$0.28
Cache read	$0.5	$0.
Cache write (5 min.)	$6.25	Not listed
Cache write (1 hr.)	$10	Not listed
Price verified	07/30/2026 Source	07/30/2026 Source

Claude Opus 5DeepSeek-V4-Flash

API inputUSD per 1M tokens, base tier

Claude Opus 5$5

DeepSeek-V4-Flash$0.14

API outputUSD per 1M tokens, base tier

Claude Opus 5$25

DeepSeek-V4-Flash$0.28

Cache readUSD per 1M tokens, base tier

Claude Opus 5$0.5

DeepSeek-V4-Flash$0.003

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 Claude Opus 5 and DeepSeek-V4-Flash
Model	Provider	Purchase route and tariff	Input per 1M tokens	Output per 1M tokens	Source
Claude Opus 5	Anthropic	Direct from developerStandard API	Lowest listed price: $5	Lowest listed price: $25	Source07/30/2026
	Amazon Bedrock	Via OpenRouteramazon-bedrock	Lowest listed price: $5	Lowest listed price: $25	Source09/07/2026
	Amazon Bedrock	Via OpenRouteramazon-bedrock/us-east-1	$5.5	$27.5	Source09/07/2026
	Anthropic	Via OpenRouteranthropic	Lowest listed price: $5	Lowest listed price: $25	Source09/07/2026
	Anthropic	Via OpenRouteranthropic/fast	$10	$50	Source09/07/2026
	Azure	Via OpenRouterazure/global	Lowest listed price: $5	Lowest listed price: $25	Source09/07/2026
	Azure	Via OpenRouterazure/us	Lowest listed price: $5	Lowest listed price: $25	Source09/07/2026
	Claude Platform on AWS	Via OpenRouterclaude-on-aws	Lowest listed price: $5	Lowest listed price: $25	Source09/07/2026
	Google	Via OpenRoutergoogle-vertex/europe	$5.5	$27.5	Source09/07/2026
	Google	Via OpenRoutergoogle-vertex/global	Lowest listed price: $5	Lowest listed price: $25	Source09/07/2026
	Google	Via OpenRoutergoogle-vertex/us	$5.5	$27.5	Source09/07/2026
DeepSeek-V4-Flash	DeepSeek	Direct from developerStandard API	$0.14	$0.28	Source07/30/2026
	Alibaba	Via OpenRouteralibaba	$0.352	$1.056	Source09/07/2026
	AtlasCloud	Via OpenRouteratlas-cloud/fp4	$0.44	$1.32	Source09/07/2026
	Baidu	Via OpenRouterbaidu/fp8	$0.14	$0.28	Source09/07/2026
	BaseTen	Via OpenRouterbaseten/fp8#1	$0.13	$0.26	Source09/07/2026
	BaseTen	Via OpenRouterbaseten/fp8#2	$0.13	$0.26	Source09/07/2026
	Cloudflare	Via OpenRoutercloudflare	$0.44	$1.32	Source09/07/2026
	CoreWeave	Via OpenRoutercoreweave/fp8	$0.13	$0.28	Source09/07/2026
	DeepInfra	Via OpenRouterdeepinfra/fp8	$0.06	$0.18	Source09/07/2026
	DeepSeek	Via OpenRouterdeepseek	$0.22	$0.66	Source09/07/2026
	DigitalOcean	Via OpenRouterdigitalocean	$0.08	$0.252	Source09/07/2026
	Fireworks	Via OpenRouterfireworks	$0.22	$0.66	Source09/07/2026
	GMICloud	Via OpenRoutergmicloud/fp8	$0.352	$1.056	Source09/07/2026
	Inceptron	Via OpenRouterinceptron/fp4	$0.13	$0.28	Source09/07/2026
	Makora	Via OpenRoutermakora	$0.09	$0.195	Source09/07/2026
	Mancer 2	Via OpenRoutermancer/fp8	$0.15	$0.5	Source09/07/2026
	Morph	Via OpenRoutermorph/bf16	$0.1234	$0.3475	Source09/07/2026
	NextBit	Via OpenRouternextbit/fp8	$0.4	$1.2	Source09/07/2026
	Novita	Via OpenRouternovita/fp8	$0.4092	$1.2276	Source09/07/2026
	OpenInference	Via OpenRouteropen-inference/fp8	$0.05	$0.16	Source09/07/2026
	Parasail	Via OpenRouterparasail/fp8	$0.14	$0.28	Source09/07/2026
	Phala	Via OpenRouterphala	$0.44	$1.32	Source09/07/2026
	Reka	Via OpenRouterreka/fp4	$0.11	$0.66	Source09/07/2026
	Relace	Via OpenRouterrelace/fp4	Lowest listed price: $0.045	Lowest listed price: $0.09	Source09/07/2026
	Sail Research	Via OpenRoutersail-research/fp4	$0.065	$0.18	Source09/07/2026
	SiliconFlow	Via OpenRoutersiliconflow/fp8	$0.22	$0.66	Source09/07/2026
	StreamLake	Via OpenRouterstreamlake/fp8	$0.088	$0.264	Source09/07/2026
	Together	Via OpenRoutertogether	$0.14	$0.28	Source09/07/2026
	Venice	Via OpenRoutervenice	$0.175	$0.35	Source09/07/2026
	Wafer	Via OpenRouterwafer/fast	$0.1	$0.25	Source09/07/2026

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.

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.

Claude Opus 5

DeepSeek-V4-Flash

Coding: 100 / 96.5 (1 benchmark family)

Reasoning: 100 / 96.1 (1 benchmark family)

Mathematics: 100 / 94.2 (1 benchmark family)

Knowledge: 100 / 96.1 (1 benchmark family)

Finance: 100 / 96.4 (1 benchmark family)

Legal: 100 / 96.9 (1 benchmark family)

Cybersecurity: 53.7 / 96.3 (1 benchmark family)

Health: 100 / 97.5 (1 benchmark family)

33 matched rows from 3 benchmark families were considered. The radar shows 8 of 9 comparable categories.

Claude Opus 5DeepSeek-V4-Flash

AlmanBench v0.1Accepted cases. Higher is better. No documented setup difference

Claude Opus 594.947 %

DeepSeek-V4-Flash93.003 %

CyberBenchOverall. Higher is better. No documented setup difference

Claude Opus 540.678 %

DeepSeek-V4-Flash66.716 %

LMArena Text, Style ControlOverall. Higher is better. No documented setup difference

Claude Opus 51,492.042

DeepSeek-V4-Flash1,438.462

CyberBenchPatch. Higher is better. No documented setup difference

Claude Opus 581.356 %

DeepSeek-V4-Flash72.414 %

LMArena Text, Style ControlBusiness, management and finance. Higher is better. No documented setup difference

Claude Opus 51,487.981

DeepSeek-V4-Flash1,434.968

CyberBenchPoC. Higher is better. No documented setup difference

Claude Opus 50 %

DeepSeek-V4-Flash61.017 %

LMArena Text, Style ControlChinese. Higher is better. No documented setup difference

Claude Opus 51,560.338

DeepSeek-V4-Flash1,479.62

LMArena Text, Style ControlCoding. Higher is better. No documented setup difference

Claude Opus 51,533.034

DeepSeek-V4-Flash1,479.449

LMArena Text, Style ControlCreative writing. Higher is better. No documented setup difference

Claude Opus 51,473.728

DeepSeek-V4-Flash1,408.903

LMArena Text, Style ControlEnglish. Higher is better. No documented setup difference

Claude Opus 51,491.814

DeepSeek-V4-Flash1,449.894

LMArena Text, Style ControlEntertainment, sports and media. Higher is better. No documented setup difference

Claude Opus 51,465.034

DeepSeek-V4-Flash1,406.152

LMArena Text, Style ControlExcluding ties. Higher is better. No documented setup difference

Claude Opus 51,506.469

DeepSeek-V4-Flash1,435.7

Key to the marks:better valuebetter, but not conclusivetie

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.

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

Compare Claude Opus 5 and DeepSeek-V4-Flash with other leading models using the same data and benchmark logic.

View the complete LLM comparison

Benchmark scores for Claude Opus 5 and DeepSeek-V4-Flash
Benchmark and task	Claude Opus 5	DeepSeek-V4-Flash	Source
AlmanBench v0.1Accepted cases Test details No documented difference in the test setupVersion: v0.1, results commit f34ddbfMetric: accepted casesScoring: Higher is betterSettings: Reasoning: max1,029 public tasks with one direct run per row. The target is the Alman language specification.	Winner94.947 %	93.003 %	Alman Institut Source details Data as of: 08/05/2026
CyberBenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	40.678 %	Winner66.716 %	Vals AI Source details Data as of: 09/01/2026
LMArena Text, Style ControlOverall Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 34,617 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 48,705 anonymous pairwise votes.	Winner1,492.042 95% confidence interval: ±4.567The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,438.462 95% confidence interval: ±4.21The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
CyberBenchPatch Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner81.356 %	72.414 %	Vals AI Source details Data as of: 08/03/2026
LMArena Text, Style ControlBusiness, management and finance Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 6,720 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 10,060 anonymous pairwise votes.	Winner1,487.981 95% confidence interval: ±8.263The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,434.968 95% confidence interval: ±7.142The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
CyberBenchPoC Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	0 %	Winner61.017 %	Vals AI Source details Data as of: 08/03/2026
LMArena Text, Style ControlChinese Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 2,433 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,509 anonymous pairwise votes.	Winner1,560.338 95% confidence interval: ±13.197The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,479.62 95% confidence interval: ±12.459The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlCoding Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 9,211 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 14,310 anonymous pairwise votes.	Winner1,533.034 95% confidence interval: ±7.486The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,479.449 95% confidence interval: ±6.409The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlCreative writing Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 7,531 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 8,029 anonymous pairwise votes.	Winner1,473.728 95% confidence interval: ±8.228The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,408.903 95% confidence interval: ±7.781The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlEnglish Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 14,349 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 21,747 anonymous pairwise votes.	Winner1,491.814 95% confidence interval: ±6.263The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,449.894 95% confidence interval: ±5.421The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlEntertainment, sports and media Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 9,181 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 10,655 anonymous pairwise votes.	Winner1,465.034 95% confidence interval: ±7.594The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,406.152 95% confidence interval: ±7.06The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlExcluding ties Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 26,397 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 37,039 anonymous pairwise votes.	Winner1,506.469 95% confidence interval: ±6.027The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,435.7 95% confidence interval: ±5.462The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlExpert prompts Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 3,824 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 4,907 anonymous pairwise votes.	Winner1,539.803 95% confidence interval: ±10.626The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,467.744 95% confidence interval: ±9.211The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlFrench Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 1,139 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,872 anonymous pairwise votes.	Winner1,508.554 95% confidence interval: ±19.274The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,457.512 95% confidence interval: ±15.76The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlGerman Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 546 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 881 anonymous pairwise votes.	Winner1,492.265 95% confidence interval: ±26.41The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,434.001 95% confidence interval: ±20.992The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlHard prompts Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 22,861 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 32,479 anonymous pairwise votes.	Winner1,517.361 95% confidence interval: ±5.496The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,458.492 95% confidence interval: ±4.934The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlHard prompts, English Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 9,347 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 15,135 anonymous pairwise votes.	Winner1,514.322 95% confidence interval: ±7.327The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,465.194 95% confidence interval: ±6.261The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlInstruction following Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 12,288 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 16,813 anonymous pairwise votes.	Winner1,497.4 95% confidence interval: ±6.794The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,433.908 95% confidence interval: ±6.024The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlJapanese Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 551 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 600 anonymous pairwise votes.	Winner1,500.289 95% confidence interval: ±28.154The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,422.217 95% confidence interval: ±25.99The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlKorean Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 685 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 909 anonymous pairwise votes.	Winner1,495.177 95% confidence interval: ±24.664The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,374.943 95% confidence interval: ±21.551The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlLegal and government Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 2,913 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 3,872 anonymous pairwise votes.	Winner1,505.46 95% confidence interval: ±11.934The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,459.091 95% confidence interval: ±10.242The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlLife, physical and social science Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 6,030 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 7,968 anonymous pairwise votes.	Winner1,512.989 95% confidence interval: ±8.565The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,460.882 95% confidence interval: ±7.588The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlLonger queries Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 16,621 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 21,664 anonymous pairwise votes.	Winner1,507.118 95% confidence interval: ±6.336The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,449.036 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.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlMath Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 1,500 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,459 anonymous pairwise votes.	Winner1,524.92 95% confidence interval: ±16.085The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,441.301 95% confidence interval: ±12.311The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlMathematical professions Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 1,813 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 2,607 anonymous pairwise votes.	Winner1,536.772 95% confidence interval: ±15.115The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,443.348 95% confidence interval: ±12.313The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlMedicine and healthcare Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 2,671 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 3,495 anonymous pairwise votes.	Winner1,500.739 95% confidence interval: ±12.393The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,462.945 95% confidence interval: ±10.866The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlMulti-turn Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 5,554 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 8,674 anonymous pairwise votes.	Winner1,485.325 95% confidence interval: ±8.848The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,445.683 95% confidence interval: ±7.508The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlNon-English Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 20,268 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 26,956 anonymous pairwise votes.	Winner1,486.228 95% confidence interval: ±5.636The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,422.651 95% confidence interval: ±5.123The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlPolish Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 532 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,014 anonymous pairwise votes.	Winner1,501.791 95% confidence interval: ±26.247The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,430.358 95% confidence interval: ±18.866The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlRussian Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 3,748 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 5,207 anonymous pairwise votes.	Winner1,501.362 95% confidence interval: ±10.435The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,437.38 95% confidence interval: ±8.983The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlSoftware and IT services Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 13,210 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 20,031 anonymous pairwise votes.	Winner1,525.012 95% confidence interval: ±6.601The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,473.269 95% confidence interval: ±5.664The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlSpanish Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 952 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 1,443 anonymous pairwise votes.	Winner1,473.649 95% confidence interval: ±20.401The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,431.772 95% confidence interval: ±16.888The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
LMArena Text, Style ControlWriting, literature and language Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Opus 5: Snapshot: claude-opus-5-high, Reasoning: high, Harness: Arena text, style control DeepSeek-V4-Flash: Snapshot: deepseek-v4-flash-high-preview, Reasoning: high, Harness: Arena text, style controlClaude Opus 5: Style-controlled Arena rating from 9,688 anonymous pairwise votes.DeepSeek-V4-Flash: Style-controlled Arena rating from 11,861 anonymous pairwise votes.	Winner1,488.93 95% confidence interval: ±7.393The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	1,421.395 95% confidence interval: ±6.717The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
Key to the marks:better valuebetter, but not conclusivetie