DeepSeek Open weights: deepseek-ai/DeepSeek-V4.1-Flash-1017Reasoning modelNewReleased 17 Oct 2026Compare with another model

DeepSeek V4.1 Flash 1017

Reasoning effort

DeepSeek V4.1 Flash 1017 is a DeepSeek open-weights reasoning model, released 17 Oct 2026. Its best configuration (max reasoning effort) ranks #72 of 775 on the BenchLeader Index at 63.1 ±6.8, in the upper half. The ± is the point: 207 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (81) and lowest in knowledge (64). At $0.525 per million tokens blended it is mid-priced. Output speed of 209 tokens per second puts it in the fastest quarter, with a first answer in 10.7 s. It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 17 Oct 2026.

Blended price: $0.525/M; $0.300 in · $1.20 out
Output speed: 209 tok/s; measured by Artificial Analysis
First answer: 11 s; first token 0.86 s
Context: 1.0M
Full answer: 13 s; median, reasoning included
Cost per run: $0.262; one full Intelligence Index run
Released: 17 Oct 2026

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index63
Reasoning81
Knowledge64
Multimodal64
Long context67
Composite80

Versions

DeepSeek has shipped 2 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.

Model	Released	Index	Rank
NewDeepSeek V4.1 Flash 1017maxthis page	17 Oct 2026	63.1	#72
DeepSeek V4.1 Flashmax	10 Sept 2026	62.0	#90

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
maxbest	63.1	#72	209 tok/s	11 s	$0.0005	Composite 80 · Knowledge 64 · Long context 67 · Multimodal 64 · Reasoning 81
not stated	–	–	60 tok/s	0.86 s	$0.0005	Composite 63

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	max	not stated	Source
LiveBench 2.0 Reasoningnot in indexeffort not stated	–	89.1%#19	LiveBench
Humanity's Last Exam (AA)not in index	43.3%#58	–	Artificial Analysis
CritPt	24.0%#48	–	Artificial Analysis

Coding

Benchmark	max	not stated	Source
LiveBench 2.0 Codingnot in indexeffort not stated	–	96.0%#4	LiveBench
SciCode (AA)not in index	55.9%#42	–	Artificial Analysis
Code Migrationnot in index	44.2%#24	–	Vals AI
Vibe Code Bench v1.1not in index	87.7%#11	–	Vals AI
Terminal-Bench 4.0 (AA)not in index	33.8%#37	–	Artificial Analysis
FrontierCode 1.1 (AA)not in index	43.7%#14	–	Artificial Analysis

Agents & tools

Benchmark	max	not stated	Source
LiveBench Long Agenticnot in indexeffort not stated	–	50.9%#12	LiveBench
LiveBench Browsernot in indexeffort not stated	–	76.0%#19	LiveBench
GDPval-AA v2.1not in index	62.3%#7	–	Artificial Analysis
Finance Agent v2not in index	57.0%#17	–	Vals AI
Harvey's Legal Agent Benchmarknot in index	14.2%#7	–	Vals AI
Legal Research Benchnot in index	45.2%#17	–	Vals AI
Tax Agent Benchnot in index	44.4%#5	–	Vals AI
Terminal-Bench 4.0 (Vals)not in index	31.8%#17	–	Vals AI
AutomationBenchnot in index	66.6%#12	–	Artificial Analysis
GDP.pdf (AA)not in index	17.2%#82	–	Artificial Analysis
AA-Briefcase v1.1not in index	1570#18	–	Artificial Analysis
Terminal-Bench-Science 0.1 (AA)not in index	12.9%#22	–	Artificial Analysis

Knowledge

Benchmark	max	not stated	Source
LiveBench 2.0 Data Analysisnot in indexeffort not stated	–	79.5%#12	LiveBench
LiveBench Researchnot in indexeffort not stated	–	79.5%#18	LiveBench
AA-Omniscience	4.7#108	–	Artificial Analysis
Excel Modeling Benchmarknot in index	61.8%#32	–	Vals AI
AA-Omniscience: accuracynot in index	42.9%#103	–	Artificial Analysis
AA-Omniscience: non-hallucinationnot in index	33.3%#200	–	Artificial Analysis

Multimodal

Benchmark	max	not stated	Source	Trend
LiveBench Designnot in indexeffort not stated	–	83.4%#4	LiveBench
MMMU-Pro	80.6%#44	–	Artificial Analysis

Long context

Benchmark	max	Source	Trend
AA-LCR	82.7%#35	Artificial Analysis

Composite

Benchmark	max	not stated	Source
LiveBench 2.0effort not stated	–	79.2%#10	LiveBench
AA Intelligence Index v4.3.2	45.1#36	–	Artificial Analysis
Vals Indexnot in index	61.3#8	–	Vals AI

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

Provider	Speed	First token	Input $/M	Output $/M	Context	Quantisation
Baseten	228 tok/s	0.37 s	$0.300	$1.20	1.0M	–
Together	227 tok/s	0.23 s	$0.300	$1.20	1.0M	–
Wafer	141 tok/s	0.61 s	$0.160	$0.500	1.0M	mxfp4
inference.net	124 tok/s	0.53 s	$0.080	$0.500	1.0M	fp8
AtlasCloud	121 tok/s	1.06 s	$0.300	$1.20	1.0M	fp8
inference.net	117 tok/s	0.44 s	$0.080	$0.500	1.0M	fp8
Morph	116 tok/s	0.76 s	$0.172	$0.600	1.0M	fp8
Parasail	114 tok/s	0.80 s	$0.300	$1.20	1.0M	fp8
SiliconFlow	111 tok/s	1.50 s	$0.300	$1.20	1.0M	fp8
Prime Intellect	103 tok/s	1.49 s	$0.300	$1.20	1.0M	–
GMICloud	103 tok/s	1.78 s	$0.300	$1.20	1.0M	fp8
Relace	100 tok/s	0.86 s	$0.070	$0.500	1.0M	–
Fireworks	98 tok/s	1.85 s	$0.300	$1.20	1.0M	–
Crusoe	92 tok/s	0.79 s	$0.290	$1.20	1.0M	fp8
DeepInfra	77 tok/s	0.58 s	$0.098	$0.294	1.0M	fp8
Decart	77 tok/s	1.63 s	$0.090	$0.180	1.0M	fp4
DeepInfra	73 tok/s	0.93 s	$0.147	$0.441	1.0M	fp8
DeepInfra	67 tok/s	1.95 s	$0.078	$0.235	1.0M	fp8
DeepSeek	64 tok/s	0.86 s	$0.150	$0.600	1.0M	–
Alibaba Cloud Int.	58 tok/s	0.79 s	$0.330	$1.32	1M	–
NovitaAI	56 tok/s	3.60 s	$0.300	$1.20	1.0M	fp8
Cloudflare	47 tok/s	6.05 s	$0.100	$0.600	1.0M	–

Price history

Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.

input outputnow $0.049 in · $0.600 out

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.006 per 1M. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.0005	$0.0004	12.1 s
Summarize a 30-page report	12,000 / 600	$0.0043	$0.0017	13.5 s
Code edit	6,000 / 1,500	$0.0036	$0.0023	17.8 s
Agentic coding session	60,000 / 4,000	$0.023	$0.0096	29.8 s
Structured extraction	2,000 / 200	$0.0008	$0.0004	11.6 s