Effort: max · Max
deepseekReleased Sep 10, 2026
DeepSeek: DeepSeek V4.1 Flash
Causal encoder-decoder activates 8B parameters for input and 16B for output, reducing the compute needed for long inputs.
Source description
DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...
Context window1M
Max output393.2K
Provider endpoints45
Benchmark results19
Measured performance
All leaderboards ↗Benchmark results
Effort: max · Max
GDPval-AA ↗2.1
1,600Effort: max · Max
AutomationBench-AA ↗AA Index 4.3.2 collection
68.89%Effort: max · Max
Effort: max · Max · Harness unreported
SciCode ↗AA Index 4.3.2 collection
51.85%AA · Under reviewEffort: max · Max
Humanity’s Last Exam ↗AA Index 4.3.2 collection
39.25%Effort: max · Max
GDP.pdf ↗AA Index 4.3.2 collection
12.8%Effort: max · Max
CritPt ↗AA Index 4.3.2 collection
14.29%AA · Under reviewEffort: max · Max
AA-Omniscience ↗AA Index 4.3.2 collection
-5.3Effort: max · Max
AA-LCR ↗1.1
84%Effort: max · Max
LiveBench Overall ↗2.0 · 2026-10-14 (dataset 2026-09-02)
76.79%Effort: max · single run · equal mean of 7 suites
LiveBench Design ↗2.0 · 2026-10-14 (dataset 2026-09-02)
77%Effort: max · single run
LiveBench Research ↗2.0 · 2026-10-14 (dataset 2026-09-02)
85.3%Effort: max · single run
LiveBench Browser Use ↗2.0 · 2026-10-14 (dataset 2026-09-02)
80%Effort: max · single run
LiveBench Coding ↗2.0 · 2026-10-14 (dataset 2026-09-02)
85.4%Effort: max · single run
LiveBench Agentic ↗2.0 · 2026-10-14 (dataset 2026-09-02)
43.8%Effort: max · single run
LiveBench Data Analysis ↗2.0 · 2026-10-14 (dataset 2026-09-02)
79.3%Effort: max · single run
LiveBench Reasoning ↗2.0 · 2026-10-14 (dataset 2026-09-02)
86.7%Effort: max · single run
Provider measurements
1 configurationSpeed
≈10K input tokens1 requestmedian / 72h
Artificial Analysis · first-party providers · serving endpoint not reported
DeepSeek
DeepSeek V4.1 Flash (Max)
Reasoning max
- Output rate
- 175.853 tok/s
- Time to first chunk
- 1.23437 s
Last checked Source ↗
Methodology 2.2.0 & measurement details
First chunk includes request processing and network time, and may include reasoning.
- Prompt
- long · ≈10,000 input tokens
- Parallel requests
- 1
- Minimum answer
- 1,500 tokens
- Aggregation
- median over 72 hours
- Token counting
- o200k_base
- Benchmark client
- Google Cloud us-central1-a
- Serving endpoint
- Not reported
- Serving region
- Not reported
- Sample size
- Not reported
- Window end
- Not reported
Access & cost
55 quotesPrices across providers
Your token mix
USD · listed token ratesOpenRouter
Catalog quote- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
novita
via Hugging Face- Input
- $0.30
- Output
- $1.20
fireworks-ai
via Hugging Face- Input
- —
- Output
- —
featherless-ai
via Hugging Face- Input
- —
- Output
- —
deepinfra
via Hugging Face- Input
- $0.20
- Output
- $0.60
- Input
- $0.20
- Output
- $0.60
Experiential
Standard list rate
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
Catalog list rate. Plan allowances and temporary discounts are separate.
Together AI
Serverless · quantization=FP8
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
Organization: DeepSeek
Fireworks AI
Standard · service_tier=default
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
Fireworks AI
Standard · service_tier=priority
- Input
- $0.375
- Output
- $1.50
- Cached input
- $0.0075
DeepSeek
Off-peak · All times outside Mon–Fri 01:00–04:00 and 06:00–10:00 UTC; all weekends and Chinese public holidays
Conditional rate- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
Thinking and non-thinking modes; cached input uses the cache-hit rate. Token prices exclude Search/Fetch tool fees. Server-side Search/Fetch calls through DeepSeek Harness, Claude Code or Codex become billable on 2026-10-31 at 16:00 UTC (November 1, 00:00 Beijing).
DeepSeek
Peak · Mon–Fri 01:00–04:00 and 06:00–10:00 UTC, excluding Chinese public holidays
Conditional rate- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
Thinking and non-thinking modes; cached input uses the cache-hit rate. Token prices exclude Search/Fetch tool fees. Server-side Search/Fetch calls through DeepSeek Harness, Claude Code or Codex become billable on 2026-10-31 at 16:00 UTC (November 1, 00:00 Beijing).
- Input
- $0.09
- Output
- $0.18
- Cached input
- $0.018
- Input
- $0.09
- Output
- $0.30
- Cached input
- $0.006
- Input
- $0.09
- Output
- $0.55
- Cached input
- $0.006
- Input
- $0.10
- Output
- $0.65
- Cached input
- $0.06
- Input
- $0.103
- Output
- $0.715
- Cached input
- $0.008
- Input
- $0.112
- Output
- $0.336
- Cached input
- $0.00336
Source discount 30%; rates as published
- Input
- $0.135
- Output
- $0.54
- Cached input
- $0.0135
- Input
- $0.14
- Output
- $0.42
- Cached input
- $0.0042
Source discount 30%; rates as published
- Input
- $0.1497
- Output
- $0.5988
- Cached input
- $0.002994
Source discount 50.1%; rates as published
- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.015
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.03
A rate applies when all its listed conditions match. Dates and times use UTC.
- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.015
A rate applies when all its listed conditions match. Dates and times use UTC.
- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
A rate applies when all its listed conditions match. Dates and times use UTC.
DeepSeek
via OpenRouterdeepseek · 00:00–01:00 UTC · mon/tue/wed/thu/fri UTC · rate option 2
Conditional rate- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
A rate applies when all its listed conditions match. Dates and times use UTC.
DeepSeek
via OpenRouterdeepseek · 01:00–04:00 UTC · mon/tue/wed/thu/fri UTC · rate option 3
Conditional rate- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
A rate applies when all its listed conditions match. Dates and times use UTC.
DeepSeek
via OpenRouterdeepseek · 04:00–06:00 UTC · mon/tue/wed/thu/fri UTC · rate option 4
Conditional rate- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
A rate applies when all its listed conditions match. Dates and times use UTC.
DeepSeek
via OpenRouterdeepseek · 06:00–10:00 UTC · mon/tue/wed/thu/fri UTC · rate option 5
Conditional rate- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
A rate applies when all its listed conditions match. Dates and times use UTC.
DeepSeek
via OpenRouterdeepseek · 10:00–00:00 UTC · mon/tue/wed/thu/fri UTC · rate option 6
Conditional rate- Input
- $0.15
- Output
- $0.60
- Cached input
- $0.003
A rate applies when all its listed conditions match. Dates and times use UTC.
- Input
- $0.15
- Output
- $0.65
- Cached input
- $0.002
- Input
- $0.18
- Output
- $0.72
- Cached input
- $0.0036
Source discount 40%; rates as published
- Input
- $0.195
- Output
- $0.78
- Cached input
- $0.006
- Input
- $0.195
- Output
- $0.78
- Cached input
- $0.0039
Source discount 35%; rates as published
- Input
- $0.20
- Output
- $0.60
- Cached input
- $0.008
- Input
- $0.20
- Output
- $0.60
- Cached input
- $0.008
- Input
- $0.21
- Output
- $0.63
- Cached input
- $0.0063
Source discount 30%; rates as published
- Input
- $0.21
- Output
- $0.84
- Cached input
- $0.0042
Source discount 30%; rates as published
- Input
- $0.27
- Output
- $1.10
- Cached input
- $0.005
- Input
- $0.27
- Output
- $1.15
- Cached input
- $0.009
- Input
- $0.29
- Output
- $1.20
- Cached input
- $0.007
- Input
- $0.30
- Output
- $1.20
Endpoint degraded (status -2)
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.03
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.007
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.03
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.007
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.006
- Input
- $0.30
- Output
- $1.20
- Cached input
- $0.0075
- Input
- $0.45
- Output
- $1.80
- Cached input
- $0.009
- Input
- $0.60
- Output
- $2.40
- Cached input
- $0.014
Offers for this model
1 offerExperiential Labs · Discount
100% off · Experiential Pro model discounts
Paid Pro plan required, covering the first 2,000 credits of usage each month, then list price.
Checked Source ↗
Model specifications
- Developer
- deepseek
- Modality
- Language
- Context
- 1,048,576 tokens
- Maximum output
- 393,216 tokens
- Weights
- Open weights
- Released
- Sep 10, 2026
- Provider listed
- Sep 10, 2026
- First discovered
- Oct 16, 2026
Exact identities
- openrouter
deepseek/deepseek-v4.1-flash - huggingface
deepseek-ai/DeepSeek-V4.1-Flash - deepinfra
deepseek-ai/DeepSeek-V4.1-Flash - experiential
deepseek-v4.1-flash - artificial-analysis
dbe7c625-3100-4463-b479-a228c41f75dd - livebench
deepseek-v4.1-flash
Sources
Updated Oct 17, 2026
More from deepseek
View all ↗deepseekListed Sep 14
DeepSeek: DeepSeek Pro Latest
This model always redirects to the latest model in the DeepSeek Pro family.
1M contextTool callingStructured outputNo reported score
No reported score
deepseekListed Sep 14
DeepSeek: DeepSeek Flash Latest
This model always redirects to the latest model in the DeepSeek Flash family.
1M contextImage inputTool callingNo reported score
No reported score
deepseekListed Sep 10
DeepSeek: DeepSeek V4.1 Flash (batch)
DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...
1M contextImage inputTool callingNo reported score
No reported score
$0.05in$0.15out/ 1M tokens · USDSail Research · via OpenRouter · 3 endpointssail-research/fp4 · quantization=fp4
deepseekListed Aug 1
DeepSeek: DeepSeek V4 Flash Latest
This model always redirects to the latest model in the DeepSeek V4 Flash family.
1M contextTool callingStructured outputNo reported score
No reported score