DeepSeek

Model lab · 14 models · also sold by 45 other providers · deepseek.com

DeepSeek models

			In / out per 1M			Cheapest route
DeepSeek V4.1 Flash 1017 1MCheapest: $0.081 on Particle AI (us-flex) 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.	$0.15	$0.6	$0.15 / $0.6	$0.262	1M	$0.081 on Particle AI (us-flex)	·	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.
DeepSeek V4.1 Flash 1M 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.	$0.045	$0.3	$0.045 / $0.3	$0.109	1M	on InferenceNet (fp8)	·	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.
DeepSeek V4 Flash Vision Exp 1M 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 capabi	$0.44	$1.32	$0.44 / $1.32	$0.66	1M	on DeepInfra (fp8)	·	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 capabi
DeepSeek V4 Pro 0813 1M DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek.	$0.66	$1.98	$0.66 / $1.98	$0.99	1M	Lab list price	·	DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek.
DeepSeek V4 Flash 0731 1M DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.019	$0.3	$0.019 / $0.3	$0.089	1M	on Sail Research (us)	·	DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0423 1M DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.	$0.09	$0.18	$0.09 / $0.18	$0.113	1M	on DeepInfra (fp8)	·	DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.
DeepSeek V4 Pro 0423 1M DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window.	$0.9	$1.5	$0.9 / $1.5	$1.05	1M	on Relace (fp4)	·	DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window.
DeepSeek V3.2 164K DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.	$0.26	$0.38	$0.26 / $0.38	$0.29	164K	on DeepInfra (fp4)	·	DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance.
DeepSeek V3.2 Exp 164K DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.	$0.27	$0.41	$0.27 / $0.41	$0.305	164K	on SiliconFlow (fp8)	·	DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures.
DeepSeek V3.1 Terminus 164K DeepSeek-V3.1 Terminus is an update to DeepSeek V3.1 that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further	$0.27	$1	$0.27 / $1	$0.453	164K	on SiliconFlow (fp8)	·	DeepSeek-V3.1 Terminus is an update to DeepSeek V3.1 that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further
DeepSeek V3.1 164K DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates.	$0.25	$0.95	$0.25 / $0.95	$0.425	164K	on DeepInfra (fp4)	·	DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates.
R1 0528 164K May 28th update to the original DeepSeek R1 Performance on par with OpenAI o1, but open-sourced and with fully open reasoning tokens.	$0.5	$2.15	$0.5 / $2.15	$0.913	164K	on DeepInfra (fp4)	·	May 28th update to the original DeepSeek R1 Performance on par with OpenAI o1, but open-sourced and with fully open reasoning tokens.
DeepSeek V3 0324 164K DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team.	$0.25	$1	$0.25 / $1	$0.438	164K	on SiliconFlow (fp8)	·	DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team.
DeepSeek V3 164K DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.	$0.256	$0.712	$0.256 / $0.712	$0.37	164K	on DeepInfra (fp4-flex)	·	DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions.

On other routes

About DeepSeek

Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd., doing business as DeepSeek, is a Chinese artificial intelligence (AI) company that develops open weights large language models (LLMs).

Caching: Separate cache hit and cache miss input prices (e.g. deepseek-flash off-peak $0.003 hit vs $0.15 miss per 1M); off-peak rates are half of peak rates

DeepSeek on the price map

Across GlyphCompany profile breaches Funding rounds labs 2 leaders 2 ranked models

Model	In / out per 1M	Vs headline
DeepSeek V4.1 Flash 1017 us-flex · 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.	$0.025 / $0.25-69% vs headline	-69%	us-flex · 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.
DeepSeek V4 Flash 0731 us, fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.019 / $0.3Same vs headline	Same	us, fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0731 fp8 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.06 / $0.18+1% vs headline	+1%	fp8 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0731 fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.021 / $0.3+2% vs headline	+2%	fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0731 fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.03 / $0.32+15% vs headline	+15%	fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0731 us, nvfp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.043 / $0.3+21% vs headline	+21%	us, nvfp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4.1 Flash fp8 · 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.	$0.045 / $0.3Same vs headline	Same	fp8 · 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.
DeepSeek V4 Flash 0423 fp8 · DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.	$0.09 / $0.18Same vs headline	Same	fp8 · DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.
DeepSeek V4.1 Flash 1017 fp4-us, fp4 · 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.	$0.101 / $0.15-57% vs headline	-57%	fp4-us, fp4 · 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.
DeepSeek V4 Flash 0423 DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.	$0.098 / $0.196+9% vs headline	+9%	DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.
DeepSeek V4 Flash 0731 fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.09 / $0.23+40% vs headline	+40%	fp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0423 DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.	$0.105 / $0.21+17% vs headline	+17%	DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window.
DeepSeek V4 Flash 0731 fast, nvfp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.08 / $0.35+65% vs headline	+65%	fast, nvfp4 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
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.	$0.119 / $0.238+67% vs headline	+67%	DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4.1 Flash fp4 · 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.	$0.035 / $0.5+39% vs headline	+39%	fp4 · 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.
DeepSeek V4.1 Flash 1017 fp4 · 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.	$0.035 / $0.5-42% vs headline	-42%	fp4 · 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.
DeepSeek V4.1 Flash 1017 us, mxfp4 · 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.	$0.038 / $0.5-42% vs headline	-42%	us, mxfp4 · 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.
DeepSeek V4.1 Flash us, mxfp4 · 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.	$0.045 / $0.5+46% vs headline	+46%	us, mxfp4 · 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.
DeepSeek V4 Flash 0731 fp32 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.13 / $0.26+82% vs headline	+82%	fp32 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.
DeepSeek V4 Flash 0731 fp8 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.	$0.13 / $0.26+82% vs headline	+82%	fp8 · DeepSeek V4 Flash 0731 is a sparse mixture-of-experts model from DeepSeek, with 13B active parameters out of 284B total.