Model profile · DeepSeek

deepseek-v4-flash-thinking

Unknown weightsmid · registry tag 2026 benchmark-derived

Thin verified coverage

Reads as thin verified coverage across the resolved source data.

Visible coverage: 0.7%
Verified coverage: 0.7%
Spread: n/a
Last verified: Jun 20, 2026

textcode1 aliases27 official source links

Open compare

Data version

Current snapshot.

Data version Jun 20, 2026Model list checked9 providers · 1081 tracked modelsPage refreshed Jul 5, 2026

The registry snapshot and page stamp are shown so a stale deploy is visible at a glance.

Source-linked scores by benchmark

Each row keeps the benchmark source, source type, raw metric, and percentile inside its fair comparison set.

Thin verified coverageThis model currently reads as thin verified coverage across the resolved source data.

Chat / text18 benchmarks83.6%

Text Arena

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #49

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,436
Percentile: 85.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: overall. Source rank: #63. Votes: 28215. Organization: deepseek. License: MIT.

85.2% percentile inside its fair comparison set

1,436Raw benchmark valueCI 1,432 - 1,441

Text Arena · Creative Writing

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #47

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,409
Percentile: 85.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: creative_writing. Source rank: #61. Votes: 4492. Organization: deepseek. License: MIT.

85.8% percentile inside its fair comparison set

1,409Raw benchmark valueCI 1,399 - 1,418

Text Arena · English

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #52

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,445
Percentile: 84.3%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: english. Source rank: #64. Votes: 13001. Organization: deepseek. License: MIT.

84.3% percentile inside its fair comparison set

1,445Raw benchmark valueCI 1,439 - 1,452

Text Arena · Exclude Ties

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #49

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 85.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: exclude_ties. Source rank: #63. Votes: 21534. Organization: deepseek. License: MIT.

85.2% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,427 - 1,439

Text Arena · Hard Prompts

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #53

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,456
Percentile: 84%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: hard_prompts. Source rank: #67. Votes: 18593. Organization: deepseek. License: MIT.

84% percentile inside its fair comparison set

1,456Raw benchmark valueCI 1,451 - 1,462

Text Arena · Hard Prompts English

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #55

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,459
Percentile: 83.3%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: hard_prompts_english. Source rank: #68. Votes: 9017. Organization: deepseek. License: MIT.

83.3% percentile inside its fair comparison set

1,459Raw benchmark valueCI 1,452 - 1,466

Text Arena · Instruction Following

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #45

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,431
Percentile: 86.5%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: instruction_following. Source rank: #56. Votes: 9578. Organization: deepseek. License: MIT.

86.5% percentile inside its fair comparison set

1,431Raw benchmark valueCI 1,424 - 1,439

Text Arena · Longer Query

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #46

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,448
Percentile: 85.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: longer_query. Source rank: #57. Votes: 12107. Organization: deepseek. License: MIT.

85.2% percentile inside its fair comparison set

1,448Raw benchmark valueCI 1,441 - 1,454

Text Arena · Multi Turn

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #49

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,445
Percentile: 85.1%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: multi_turn. Source rank: #63. Votes: 5015. Organization: deepseek. License: MIT.

85.1% percentile inside its fair comparison set

1,445Raw benchmark valueCI 1,436 - 1,455

Text Arena · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #59

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,422
Percentile: 82.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: overall. Source rank: #72. Votes: 28215. Organization: deepseek. License: MIT.

82.2% percentile inside its fair comparison set

1,422Raw benchmark valueCI 1,417 - 1,427

Text Arena · Creative Writing · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #57

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,398
Percentile: 82.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: creative_writing. Source rank: #70. Votes: 4492. Organization: deepseek. License: MIT.

82.7% percentile inside its fair comparison set

1,398Raw benchmark valueCI 1,388 - 1,407

Text Arena · English · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #69

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,430
Percentile: 79.1%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: english. Source rank: #81. Votes: 13001. Organization: deepseek. License: MIT.

79.1% percentile inside its fair comparison set

1,430Raw benchmark valueCI 1,424 - 1,436

Text Arena · Exclude Ties · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #61

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,411
Percentile: 81.5%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: exclude_ties. Source rank: #74. Votes: 21534. Organization: deepseek. License: MIT.

81.5% percentile inside its fair comparison set

1,411Raw benchmark valueCI 1,404 - 1,417

Text Arena · Hard Prompts · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #60

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,430
Percentile: 81.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: hard_prompts. Source rank: #74. Votes: 18593. Organization: deepseek. License: MIT.

81.8% percentile inside its fair comparison set

1,430Raw benchmark valueCI 1,424 - 1,436

Text Arena · Hard Prompts English · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #69

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,433
Percentile: 79%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: hard_prompts_english. Source rank: #84. Votes: 9017. Organization: deepseek. License: MIT.

79% percentile inside its fair comparison set

1,433Raw benchmark valueCI 1,426 - 1,440

Text Arena · Instruction Following · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #45

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,414
Percentile: 86.5%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: instruction_following. Source rank: #57. Votes: 9578. Organization: deepseek. License: MIT.

86.5% percentile inside its fair comparison set

1,414Raw benchmark valueCI 1,407 - 1,421

Text Arena · Longer Query · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #53

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,423
Percentile: 82.9%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: longer_query. Source rank: #66. Votes: 12107. Organization: deepseek. License: MIT.

82.9% percentile inside its fair comparison set

1,423Raw benchmark valueCI 1,417 - 1,430

Text Arena · Multi Turn · No Style Control

AR · Chat / text · Human

It tests whether the model is actually useful in normal conversational turns, not just on narrow correctness tasks.

Rank #52

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,430
Percentile: 84.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: multi_turn. Source rank: #66. Votes: 5015. Organization: deepseek. License: MIT.

84.2% percentile inside its fair comparison set

1,430Raw benchmark valueCI 1,421 - 1,439

Coding2 benchmarks80.9%

Text Arena · Coding

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #58

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,477
Percentile: 82.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: coding. Source rank: #72. Votes: 8128. Organization: deepseek. License: MIT.

82.2% percentile inside its fair comparison set

1,477Raw benchmark valueCI 1,470 - 1,485

Text Arena · Coding · No Style Control

AR · Coding · Human

It tells you whether the model can generate, repair, and reason over code under evaluator pressure rather than marketing examples.

Rank #66

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,437
Percentile: 79.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: coding. Source rank: #81. Votes: 8128. Organization: deepseek. License: MIT.

79.7% percentile inside its fair comparison set

1,437Raw benchmark valueCI 1,430 - 1,445

Reasoning / math / science2 benchmarks81.2%

Text Arena · Math

AR · Reasoning / math / science · Human

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #49

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,436
Percentile: 84.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: math. Source rank: #60. Votes: 1511. Organization: deepseek. License: MIT.

84.7% percentile inside its fair comparison set

1,436Raw benchmark valueCI 1,420 - 1,451

Text Arena · Math · No Style Control

AR · Reasoning / math / science · Human

It is one of the cleaner reads on deliberate reasoning strength rather than style or popularity.

Rank #71

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,420
Percentile: 77.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: math. Source rank: #85. Votes: 1511. Organization: deepseek. License: MIT.

77.7% percentile inside its fair comparison set

1,420Raw benchmark valueCI 1,404 - 1,435

Professional reasoning18 benchmarks82.9%

Text Arena · Expert

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena expert leaderboard.

Rank #46

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,467
Percentile: 83.6%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: expert. Source rank: #57. Votes: 2676. Organization: deepseek. License: MIT.

83.6% percentile inside its fair comparison set

1,467Raw benchmark valueCI 1,454 - 1,479

Text Arena · Industry Business And Management And Financial Operations

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_business_and_management_and_financial_operations leaderboard.

Rank #52

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,435
Percentile: 84%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_business_and_management_and_financial_operations. Source rank: #68. Votes: 5768. Organization: deepseek. License: MIT.

84% percentile inside its fair comparison set

1,435Raw benchmark valueCI 1,426 - 1,443

Text Arena · Industry Entertainment And Sports And Media

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_entertainment_and_sports_and_media leaderboard.

Rank #50

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,406
Percentile: 84.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_entertainment_and_sports_and_media. Source rank: #63. Votes: 5863. Organization: deepseek. License: MIT.

84.8% percentile inside its fair comparison set

1,406Raw benchmark valueCI 1,397 - 1,415

Text Arena · Industry Legal And Government

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_legal_and_government leaderboard.

Rank #33

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,462
Percentile: 89.3%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_legal_and_government. Source rank: #44. Votes: 2186. Organization: deepseek. License: MIT.

89.3% percentile inside its fair comparison set

1,462Raw benchmark valueCI 1,448 - 1,475

Text Arena · Industry Life And Physical And Social Science

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_life_and_physical_and_social_science leaderboard.

Rank #46

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,458
Percentile: 86.1%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_life_and_physical_and_social_science. Source rank: #58. Votes: 4689. Organization: deepseek. License: MIT.

86.1% percentile inside its fair comparison set

1,458Raw benchmark valueCI 1,449 - 1,468

Text Arena · Industry Mathematical

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_mathematical leaderboard.

Rank #51

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,443
Percentile: 83.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_mathematical. Source rank: #63. Votes: 1576. Organization: deepseek. License: MIT.

83.8% percentile inside its fair comparison set

1,443Raw benchmark valueCI 1,427 - 1,459

Text Arena · Industry Medicine And Healthcare

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_medicine_and_healthcare leaderboard.

Rank #51

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,459
Percentile: 83.1%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_medicine_and_healthcare. Source rank: #65. Votes: 2004. Organization: deepseek. License: MIT.

83.1% percentile inside its fair comparison set

1,459Raw benchmark valueCI 1,445 - 1,473

Text Arena · Industry Software And It Services

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_software_and_it_services leaderboard.

Rank #52

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,471
Percentile: 84.3%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_software_and_it_services. Source rank: #66. Votes: 11464. Organization: deepseek. License: MIT.

84.3% percentile inside its fair comparison set

1,471Raw benchmark valueCI 1,464 - 1,478

Text Arena · Industry Writing And Literature And Language

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_writing_and_literature_and_language leaderboard.

Rank #46

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,420
Percentile: 86.1%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_writing_and_literature_and_language. Source rank: #59. Votes: 6809. Organization: deepseek. License: MIT.

86.1% percentile inside its fair comparison set

1,420Raw benchmark valueCI 1,412 - 1,428

Text Arena · Expert · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena expert leaderboard.

Rank #58

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,435
Percentile: 79.3%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: expert. Source rank: #69. Votes: 2676. Organization: deepseek. License: MIT.

79.3% percentile inside its fair comparison set

1,435Raw benchmark valueCI 1,423 - 1,447

Text Arena · Industry Business And Management And Financial Operations · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_business_and_management_and_financial_operations leaderboard.

Rank #67

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,411
Percentile: 79.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_business_and_management_and_financial_operations. Source rank: #81. Votes: 5768. Organization: deepseek. License: MIT.

79.2% percentile inside its fair comparison set

1,411Raw benchmark valueCI 1,402 - 1,419

Text Arena · Industry Entertainment And Sports And Media · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_entertainment_and_sports_and_media leaderboard.

Rank #61

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,393
Percentile: 81.4%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_entertainment_and_sports_and_media. Source rank: #74. Votes: 5863. Organization: deepseek. License: MIT.

81.4% percentile inside its fair comparison set

1,393Raw benchmark valueCI 1,384 - 1,401

Text Arena · Industry Legal And Government · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_legal_and_government leaderboard.

Rank #45

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,442
Percentile: 85.2%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_legal_and_government. Source rank: #54. Votes: 2186. Organization: deepseek. License: MIT.

85.2% percentile inside its fair comparison set

1,442Raw benchmark valueCI 1,429 - 1,456

Text Arena · Industry Life And Physical And Social Science · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_life_and_physical_and_social_science leaderboard.

Rank #60

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,437
Percentile: 81.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_life_and_physical_and_social_science. Source rank: #73. Votes: 4689. Organization: deepseek. License: MIT.

81.7% percentile inside its fair comparison set

1,437Raw benchmark valueCI 1,428 - 1,446

Text Arena · Industry Mathematical · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_mathematical leaderboard.

Rank #69

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,424
Percentile: 77.9%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_mathematical. Source rank: #82. Votes: 1576. Organization: deepseek. License: MIT.

77.9% percentile inside its fair comparison set

1,424Raw benchmark valueCI 1,408 - 1,440

Text Arena · Industry Medicine And Healthcare · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_medicine_and_healthcare leaderboard.

Rank #64

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,437
Percentile: 78.6%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_medicine_and_healthcare. Source rank: #75. Votes: 2004. Organization: deepseek. License: MIT.

78.6% percentile inside its fair comparison set

1,437Raw benchmark valueCI 1,423 - 1,451

Text Arena · Industry Software And It Services · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_software_and_it_services leaderboard.

Rank #64

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,441
Percentile: 80.6%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_software_and_it_services. Source rank: #77. Votes: 11464. Organization: deepseek. License: MIT.

80.6% percentile inside its fair comparison set

1,441Raw benchmark valueCI 1,434 - 1,447

Text Arena · Industry Writing And Literature And Language · No Style Control

AR · Professional reasoning · Human

Observed user preference in Arena's Text Arena industry_writing_and_literature_and_language leaderboard.

Rank #54

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,407
Percentile: 83.6%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: industry_writing_and_literature_and_language. Source rank: #66. Votes: 6809. Organization: deepseek. License: MIT.

83.6% percentile inside its fair comparison set

1,407Raw benchmark valueCI 1,399 - 1,415

Multilingual14 benchmarks80.7%

Text Arena · Chinese

AR · Multilingual · Human

Observed user preference in Arena's Text Arena chinese leaderboard.

Rank #41

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,487
Percentile: 86.4%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: chinese. Source rank: #51. Votes: 1413. Organization: deepseek. License: MIT.

86.4% percentile inside its fair comparison set

1,487Raw benchmark valueCI 1,470 - 1,503

Text Arena · French

AR · Multilingual · Human

Observed user preference in Arena's Text Arena french leaderboard.

Rank #38

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,467
Percentile: 82.9%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: french. Source rank: #49. Votes: 1012. Organization: deepseek. License: MIT.

82.9% percentile inside its fair comparison set

1,467Raw benchmark valueCI 1,446 - 1,487

Text Arena · German

AR · Multilingual · Human

Observed user preference in Arena's Text Arena german leaderboard.

Rank #44

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,434
Percentile: 81.9%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: german. Source rank: #60. Votes: 455. Organization: deepseek. License: MIT.

81.9% percentile inside its fair comparison set

1,434Raw benchmark valueCI 1,405 - 1,462

Text Arena · Japanese

AR · Multilingual · Human

Observed user preference in Arena's Text Arena japanese leaderboard.

Rank #28

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,419
Percentile: 86.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: japanese. Source rank: #38. Votes: 273. Organization: deepseek. License: MIT.

86.7% percentile inside its fair comparison set

1,419Raw benchmark valueCI 1,381 - 1,456

Text Arena · Korean

AR · Multilingual · Human

Observed user preference in Arena's Text Arena korean leaderboard.

Rank #46

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,391
Percentile: 78.4%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: korean. Source rank: #59. Votes: 502. Organization: deepseek. License: MIT.

78.4% percentile inside its fair comparison set

1,391Raw benchmark valueCI 1,363 - 1,419

Text Arena · Russian

AR · Multilingual · Human

Observed user preference in Arena's Text Arena russian leaderboard.

Rank #46

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,438
Percentile: 84.4%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: russian. Source rank: #60. Votes: 3173. Organization: deepseek. License: MIT.

84.4% percentile inside its fair comparison set

1,438Raw benchmark valueCI 1,427 - 1,449

Text Arena · Spanish

AR · Multilingual · Human

Observed user preference in Arena's Text Arena spanish leaderboard.

Rank #48

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,434
Percentile: 78%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: spanish. Source rank: #60. Votes: 764. Organization: deepseek. License: MIT.

78% percentile inside its fair comparison set

1,434Raw benchmark valueCI 1,411 - 1,457

Text Arena · Chinese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena chinese leaderboard.

Rank #44

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,472
Percentile: 85.4%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: chinese. Source rank: #54. Votes: 1413. Organization: deepseek. License: MIT.

85.4% percentile inside its fair comparison set

1,472Raw benchmark valueCI 1,455 - 1,489

Text Arena · French · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena french leaderboard.

Rank #52

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,442
Percentile: 76.4%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: french. Source rank: #64. Votes: 1012. Organization: deepseek. License: MIT.

76.4% percentile inside its fair comparison set

1,442Raw benchmark valueCI 1,421 - 1,463

Text Arena · German · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena german leaderboard.

Rank #56

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,419
Percentile: 76.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: german. Source rank: #69. Votes: 455. Organization: deepseek. License: MIT.

76.8% percentile inside its fair comparison set

1,419Raw benchmark valueCI 1,390 - 1,447

Text Arena · Japanese · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena japanese leaderboard.

Rank #36

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,399
Percentile: 82.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: japanese. Source rank: #46. Votes: 273. Organization: deepseek. License: MIT.

82.8% percentile inside its fair comparison set

1,399Raw benchmark valueCI 1,362 - 1,437

Text Arena · Korean · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena korean leaderboard.

Rank #62

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,367
Percentile: 70.7%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: korean. Source rank: #75. Votes: 502. Organization: deepseek. License: MIT.

70.7% percentile inside its fair comparison set

1,367Raw benchmark valueCI 1,339 - 1,396

Text Arena · Russian · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena russian leaderboard.

Rank #45

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,426
Percentile: 84.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: russian. Source rank: #54. Votes: 3173. Organization: deepseek. License: MIT.

84.8% percentile inside its fair comparison set

1,426Raw benchmark valueCI 1,415 - 1,437

Text Arena · Spanish · No Style Control

AR · Multilingual · Human

Observed user preference in Arena's Text Arena spanish leaderboard.

Rank #57

verified runtimeexact aliasBackground only

Raw row drilldownsource row, percentile, last updated, eligibility

Source: Arena
Raw value: 1,420
Percentile: 73.8%
Last updated: recent
Eligibility: benchmark_derived_model

Parsed from Arena leaderboard dataset row `deepseek-v4-flash-thinking`. Category: spanish. Source rank: #70. Votes: 764. Organization: deepseek. License: MIT.

73.8% percentile inside its fair comparison set

1,420Raw benchmark valueCI 1,398 - 1,443

Source links and registry checks

official

Arena

Jun 20, 2026

deepseek-v4-flash-thinking

Current snapshot.

Source-linked scores by benchmark

Source links and registry checks

Loading model evidence.

deepseek-v4-flash-thinking

Current snapshot.

Source-linked scores by benchmark

Source links and registry checks