DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash

DeepSeek V4.1 Flash is an open-weight multimodal model that accepts text and images and generates text. The official model card documents a one-million-token context window, a mixture-of-experts architecture, and MIT-licensed weights.

Compare

Context window: 1M
Input: Image · Text
Output: Text
Released: 2026-09-10

Weights: Open
License: MIT

Pricing

$0.15

USD / 1,000,000 token

billing period: off_peak; peak hours: Mon-Fri 01:00-04:00 and 06:00-10:00 UTC, excluding Chinese public holidays

Cache read

$0.003

USD / 1,000,000 token

billing period: off_peak; peak hours: Mon-Fri 01:00-04:00 and 06:00-10:00 UTC, excluding Chinese public holidays

OUT

$0.60

USD / 1,000,000 token

billing period: off_peak; peak hours: Mon-Fri 01:00-04:00 and 06:00-10:00 UTC, excluding Chinese public holidays

Performance

Overall capability

56.1/100

Coding: 52.9
Reasoning: 52.2
Professional work: 55.5
Knowledge & language: 58.9


SimpleQA Verified	Knowledge & language	Epoch AI	2026-11-05
Chess Puzzles	Reasoning	Epoch AI	2026-11-05
AA-Analyst AgentDeepSeek V4.1 Flash (Max)	Professional work	Artificial Analysis	2026-11-05
APEX-Agents 1.1	Professional work	Mercor	2026-11-05
MLCR-AA（Artificial Analysis）DeepSeek V4.1 Flash (Max)	Knowledge & language	Artificial Analysis	2026-11-05
Terminal-Bench Science · AADeepSeek V4.1 Flash (Max)	Reasoning	Artificial Analysis	2026-11-05
Terminal-Bench 4.0 (Vals rerun)	Coding	Vals AI	2026-11-05
ProgramBench	Coding	Vals AI	2026-11-05

Providers

No providers found

ModelRegistry · Catalog updated 2026-11-05