DeepSeek

DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash

Model weightsReleased Sep 10, 2026MIT

ThinkingToolsVisionopen weights

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...

Run in ChatOSS Cloud Plans

Cost / 1M tokens

Very low usage

$0.022

input

$0.004

cached input

$0.40

output

Context Window

Up to 944K max output

Architecture

552B total · 32B active MoE

text + image → text

Quickstart & API

deepseek/deepseek-v4.1-flash

Open the model picker in ChatOSS Desktop (Code, Kanban, or Chat) and select or paste this model ID.

Best for & Strengths

Key Strengths

Agentic coding1M context windowUltra-low cache cost ($0.003/M)Fast throughputMIT open weights

Recommended Workloads in ChatOSS

Autonomous coding agents
Full-stack web app generation
High-volume agent loops

Available in ChatOSS Code, Kanban Orchestrate, and Chat modes via Plus, Max, or BYOK.

Verified Evaluations

Benchmarks & Frontier Comparison

How DeepSeek V4.1 Flash compares against leading open-weight and proprietary frontier models across independent agentic coding and reasoning suites.

View full Top 40 Leaderboard →

DeepSeek V4.1 FlashDeepSeek V4.1 FlashGLM 5.3Kimi K3.1 Pro

Benchmark (Metric)	DeepSeek V4.1 Flash	DeepSeek V4.1 Flash	GLM 5.3	Kimi K3.1 Pro	Claude Opus 5.5
Agentic & Software Engineering
ChatOSS Agentic Coding Index (%)	63.2%	63.2%	58.3%	65%	74.1%
SWE-Bench Verified (%)	—	91.5%	95.4%	94.8%	97%
DeepSWE v1.1 (Pass@1 %)	77%	77%	69%	73.5%	74.2%
Vibe Code Bench v1.1 (%)	87.73%	87.7%	78.1%	88.4%	90.3%
Terminal-Bench 4.0 (%)	42.3%	42.3%	42%	46.2%	65.2%
Code Migration (%)	45.62%	45.6%	44.2%	52%	66.7%
Intelligence & Token Economics
Artificial Analysis Intelligence Index (0–100)	—	39.5	44.8	55.1	57.6
Input Price (/ 1M tokens)	$0.022	$0.15	$1.40	$2.00	$4.00
Output Price (/ 1M tokens)	$0.40	$0.60	$4.40	$10.00	$20.00

ChatOSS Agentic Coding Index

ChatOSS Benchmark Audit · 4-benchmark composite mean

63.2%

DeepSWE v1.1

DeepSWE / Official Audit · Mini-SWE-Agent

77%

Vibe Code Bench v1.1

Vals.ai · OpenHands

87.73%

Terminal-Bench 4.0

Vals.ai / Artificial Analysis · Terminus 2

42.3%

Terminal-Bench 2.1

Vals.ai · Terminus 2

90.6%

Code Migration

Vals.ai · Vals evaluation harness

45.62%

AutomationBench

Artificial Analysis v4.3

67%

CyberGym (Pass@1)

Model Release Evaluation

88.1%

GPQA Diamond (Pass@1)

Benchmark Audit

90.9%

Humanity's Last Exam (Pass@1)

Artificial Analysis / Audit

43%

Active Provider Endpoints (36)

Upstream inference providers serving DeepSeek V4.1 Flash, with per-provider pricing, context limits, quantization, and 24-hour uptime.

Provider	Input / 1M	Output / 1M	Cache Read / 1M	Context	Max Output	Quantization	Uptime (24h)
Morphcache morph/fp8	$0.022	$0.40	$0.004	1M	944K	fp8	95.1%
PrimeIntellectcache primeintellect	$0.039	$1.20	$0.0012	1M	131K	—	98.4%
Wafercache wafer/us	$0.051	$0.567	$0.018	1M	944K	mxfp4	100.0%
OpenInferencecache open-inference/fp4	$0.07	$0.80	$0.006	1M	944K	fp4	100.0%
InferenceNetcache inference-net/us	$0.103	$0.60	$0.008	1M	393K	fp8	100.0%
InferenceNetcache inference-net/fp8	$0.103	$0.60	$0.008	1M	393K	fp8	100.0%
DeepInfracache deepinfra/fp8/flex	$0.112	$0.336	$0.0034	1M	393K	fp8	90.9%
Relacecache relace/fp4	$0.12	$0.50	$0.003	1M	944K	fp4	100.0%
Rekacache reka	$0.12	$0.64	$0.008	1M	393K	—	98.3%
CoreWeavecache coreweave/fp8	$0.13	$0.65	$0.002	1M	944K	fp8	99.7%
AtlasCloudcache atlas-cloud/fp8	$0.135	$0.54	$0.013	1M	393K	fp8	100.0%
DeepInfracache deepinfra/fp8	$0.14	$0.42	$0.0042	1M	393K	fp8	100.0%
StreamLakecache streamlake/fp8	$0.147	$0.588	$0.0029	1M	384K	fp8	100.0%
GMICloudcache gmicloud/fp8	$0.18	$0.72	$0.0036	1M	944K	fp8	100.0%
Wafercache wafer/mxfp4	$0.19	$0.65	$0.012	1M	944K	mxfp4	100.0%
Coherecache cohere/fp8	$0.195	$0.66	$0.003	1M	384K	fp8	100.0%
Novitacache novita/fp8	$0.195	$0.78	$0.0039	1M	393K	fp8	100.0%
DeepInfracache deepinfra/fp8/priority	$0.21	$0.63	$0.0063	1M	393K	fp8	100.0%
Fireworkscache fireworks/spot	$0.21	$0.84	$0.0042	1M	944K	—	96.9%
Alibabacache alibaba	$0.21	$0.84	$0.0042	1M	393K	—	100.0%
Mancer 2cache mancer/fp8	$0.212	$0.75	$0.05	1M	944K	fp8	99.7%
DigitalOceancache digitalocean	$0.225	$0.90	$0.006	1M	944K	—	99.9%
Makoracache makora/fp8	$0.26	$1.17	$0.0058	1M	944K	fp8	99.9%
Crusoecache crusoe/fp8	$0.29	$1.20	$0.007	1M	944K	fp8	100.0%
Googlecache google-vertex/global	$0.30	$1.20	$0.03	1M	944K	—	99.2%
Baiducache baidu/fp8	$0.30	$1.20	$0.006	1M	393K	fp8	100.0%
BaseTencache baseten/us	$0.30	$1.20	$0.007	1M	262K	—	95.8%
Togethercache together	$0.30	$1.20	$0.006	1M	944K	—	100.0%
SiliconFlowcache siliconflow/fp8	$0.30	$1.20	$0.006	1M	393K	fp8	99.9%
Modalcache modal	$0.30	$1.20	$0.03	1M	944K	—	100.0%
BaseTencache baseten	$0.30	$1.20	$0.007	1M	262K	—	97.9%
Parasailcache parasail/fp8	$0.30	$1.20	$0.006	1M	944K	fp8	100.0%
Fireworkscache fireworks	$0.30	$1.20	$0.006	1M	944K	—	99.4%
Venicecache venice/fp8	$0.30	$1.20	$0.0075	1M	262K	fp8	99.5%
Parasailcache parasail/fast	$0.45	$1.80	$0.009	1M	944K	fp8	100.0%
Fireworkscache fireworks/us	$0.45	$1.80	$0.009	1M	944K	—	99.9%

Supported API Parameters

Parameters accepted by this model when called through ChatOSS or OpenRouter.

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_pparallel_tool_callspresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_atop_ktop_logprobstop_p

Deep-Dive Benchmarks & Teardowns

Independent benchmark analysis and real-world agent billing teardowns featuring DeepSeek V4.1 Flash.

2026-10-27 · 10 min read

MiniMax M3.1 Flash vs. DeepSeek V4.1 Flash 1017: cheaper tokens, higher bill

Both list at $0.15/$0.60 per million tokens, and MiniMax wins the business-workflow benchmarks. But on vals.ai's Vibe Code Bench v1.1, DeepSeek's October 17 checkpoint scores 87.73% at $0.58 per test while M3.1 Flash scores 73.94% at $1.15 — because the cheaper token burns 66.86M of them per task. Here is the independent evidence and the cache-aware math.

Read article →

2026-09-21 · 11 min read

GLM 5.3 Flash vs. DeepSeek V4.1 Flash: which is better at agentic coding?

Both labs ship a million-token context, MIT weights, and a ~$0.15-per-million price tag. On the independent harnesses that have actually run both models — vals.ai's Vibe Code Bench and Terminal-Bench 2.1, BenchLM's Vibe Code Bench 1-100, and Artificial Analysis' v4.3 index — DeepSeek V4.1 Flash wins agentic coding, and by a wider margin than either vendor's own table suggests.

Read article →

2026-10-29 · 11 min read

MiniMax M3.1 Flash may be the best open-source orchestrator agent right now

AutomationBench is the closest public test of orchestration, and MiniMax M3.1 Flash wins it on both the independent run (73% vs 67%) and the vendors' table (62.8 vs 60.5). It also leads HLE in Artificial Analysis's run, 48% to 43% — but DeepSeek V4.1 Flash 1017 still owns coding, cache-heavy bills and the MIT licence. The honest case, including the numbers that cut against it.

Read article →

Compare Similar Models

Browse all models →

DeepSeek

DeepSeek V4.1 Flash 1017

DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the `deepseek-flash` API model name. V4.1 Flash is a sparse mixture-of-experts model...

$0.022 / $0.40Coding 63.2%

DeepSeek

DeepSeek V4 Flash Vision Exp

DeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of DeepSeek V4 Flash 0731 from DeepSeek, adding image understanding while matching the base model on text capabilities including agents,...

$0.44 / $1.32

DeepSeek

DeepSeek V4 Pro 0813

DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.

$0.111 / $4.44Coding 46.9%

DeepSeek

DeepSeek V4 Flash 0731

DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total. This re-post-trained revision is suited for coding, reasoning, and agent workflows....

$0.043 / $0.30Coding 44.2%

Anthropic

Claude Opus 5.5

Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...

$4.00 / $20.00Coding 74.1%

Moonshot AI

Kimi K3.1 Pro

Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, with a 1M-token context window. It is built for long-horizon coding, 3D...

$2.00 / $10.00Coding 65%