Vendor · Language models
DeepSeek has 23 models in the directory. The strongest is DeepSeek V4.1 Flash 1017: Rank 13 to 20 of 95 in the overall index.
DeepSeek released its open-weight V4 model family in 2026. The company lists V4-Pro at 1.6T total parameters, with 49B active per token. In September 2026, it added V4.1-Flash with native visual understanding to its API.
In the directory: 23 language models. 23 of them with open weights that can run locally.
Every model there ever was. Models with a profile lead to every benchmark, price and source.
| DeepSeek V4.1 Flash 1017 | Oct 17, 2026 | Open weights | 66Rank 13 to 20 of 95 | $0.15 / $0.60 | 178 | 1.05M | 68 |
| DeepSeek-V4.1-Flash | Sep 10, 2026 | Active | 48Rank 22 to 35 of 95 | $0.15 / $0.60 | 177 | 1M | 39 |
| DeepSeek V4 Flash Vision Exp | Aug 21, 2026 | Active | n/a | n/a | n/a | 1M | 44 |
| DeepSeek V4 Pro 0813 | Aug 13, 2026 | Active | n/a | n/a | n/a | 1M | 44 |
| DeepSeek V4 Flash 0731 | Jul 31, 2026 | Active | n/a | n/a | n/a | 1M | 46 |
| DeepSeek-V4-Flash | Apr 24, 2026 | Limited access | n/a | $0.14 / $0.28 | 119 | 1M | 65 |
| DeepSeek-V4-Pro | Apr 24, 2026 | Active | 18Rank 42 to 50 of 95 | $0.66 / $1.98 | 81 | 1M | 59 |
| DeepSeek V4 Pro 0423 | Apr 23, 2026 | Active | n/a | n/a | n/a | 1M | 14 |
| DeepSeek OCR 2 | Jan 27, 2026 | Active | n/a | n/a | n/a | 8K | 41 |
| DeepSeek-V3.2 | December 2025 | Active | below 1Rank 76 to 89 of 95 | $0.28 / $0.42 | n/a | 128K | 47 |
| DeepSeek-V3.2 Exp | Sep 29, 2025 | Active | n/a | n/a | n/a | 128K | 41 |
| DeepSeek-V3.1 Terminus | Sep 22, 2025 | Active | n/a | $0.27 / $1 | 75 | 164K | 38 |
| DeepSeek-V3.1 | Aug 21, 2025 | Active | n/a | $0.27 / $1 | 119 | 128K | 39 |
| DeepSeek R1 0528 Qwen3 8B | May 28, 2025 | Active | n/a | n/a | n/a | 131K | 60 |
| DeepSeek-R1 0528 | May 28, 2025 | Active | n/a | $0.50 / $2.18 | 38 | 164K | 38 |
| DeepSeek-V3 0324 | Mar 24, 2025 | Active | below 1Rank 86 to 93 of 95 | $0.25 / $1 | 135 | 164K | 49 |
| DeepSeek-R1-Distill-Llama-70B | Jan 20, 2025 | Active | n/a | n/a | 22 | 128K | 44 |
| DeepSeek-R1-Distill-Qwen-32B | Jan 20, 2025 | Active | n/a | n/a | n/a | 131K | 40 |
| DeepSeek-R1 | January 2025 | Active | below 1Rank 78 to 88 of 95 | $0.70 / $2.50 | n/a | 128K | 58 |
| DeepSeek-V3 | Dec 26, 2024 | Active | below 1Rank 92 to 94 of 95 | $0.26 / $0.71 | 62 | 128K | 54 |
| DeepSeek Coder V2 | June 2024 | Active | n/a | n/a | n/a | 131K | 32 |
| DeepSeek-V2no profile | May 2024 | No status data | n/a | n/a | n/a | n/a | n/a |
| DeepSeek LLM 67B | November 2023 | Active | n/a | n/a | n/a | 4K | 25 |
Language model · Oct 17, 2026
DeepSeek V4.1 Flash 1017 is the checkpoint released on October 17 under the API ID deepseek-flash. Its model card provides the weights under MIT. It accepts text and images, supports a 1,048,576-token context, and allows up to 393,216 output tokens.
Language model · Sep 10, 2026
New in DeepSeek-V4.1-Flash are native image input and asymmetric activation. DeepSeek describes 552 billion backbone parameters, 8 billion active parameters during prefill, 16 billion during decoding, and a 1,000,000-token context window. Reasoning effort can be adjusted more finely than in the original V4-Flash line.
Language model · Aug 21, 2026
DeepSeek V4 Flash Vision Exp was an experimental multimodal V4-Flash checkpoint for visual agents. Publisher metadata totals about 304.6 billion parameters and shows image-to-text processing through the V4 inference environment. DeepSeek presented it as a visual extension of the text model.
Language model · Aug 13, 2026
DeepSeek V4 Pro 0813 is the official V4-Pro release associated with the general-availability API release announced on August 13, 2026. The model card releases the weights under MIT and recommends up to 384,000 output tokens at the higher reasoning levels.
Language model · Jul 31, 2026
DeepSeek V4 Flash 0731 is the dated V4-Flash checkpoint from July 31, 2026. Its model card lists MIT, very long context, and a recommended maximum output of 384,000 tokens at the higher reasoning levels. The date separates it from the current deepseek-flash route.
Language model · Apr 24, 2026
New in DeepSeek-V4-Pro is the combination of an open V4-Pro checkpoint and a 1,000,000-token context window. DeepSeek describes 1.6 trillion total parameters and 49 billion active parameters.
Language model · Apr 23, 2026
DeepSeek V4 Pro 0423 is the catalog entry dated April 23, 2026. DeepSeek dates the public V4 announcement to April 24, so 0423 identifies an initial snapshot rather than a separately confirmed weight release. The date distinguishes it from the later DeepSeek V4 Pro 0813.
Language model · Jan 27, 2026
What is distinctive about DeepSeek OCR 2 is its document processing with dynamic resolution. The 3-billion-parameter checkpoint combines multiple crops with an overview and, according to its model card, supports free OCR and document-to-Markdown prompts.
Language model · December 2025
New in DeepSeek-V3.2 is DeepSeek Sparse Attention as a component for long context and tool use. The released design includes 671 billion parameters in the main model, 37 billion active parameters, and a separate 14-billion-parameter MTP component. Thinking and non-thinking modes remain available with a 128,000-token context window.
Language model · Sep 29, 2025
What is distinctive about DeepSeek-V3.2 Exp is its experimental introduction of DeepSeek Sparse Attention into the V3 line. The checkpoint follows DeepSeek-V3.1 Terminus, uses 671 billion total parameters and 37 billion active parameters.
Language model · Sep 22, 2025
DeepSeek-V3.1 Terminus is the stabilized V3.1 release for hybrid inference. It follows DeepSeek-V3.1 and precedes the experimental DeepSeek-V3.2 Exp. The mixture-of-experts checkpoint has 671 billion total parameters and 37 billion active parameters.
Language model · Aug 21, 2025
New in DeepSeek-V3.1 are separate thinking and non-thinking modes. The checkpoint remains a 671-billion-parameter mixture-of-experts model with 37 billion active parameters and a 128,000-token context window.
9 more current models are in the model list above.
Derived from the overall index, the API price and the measured speed of the current models.
The strongest choice
DeepSeek V4.1 Flash 1017
The vendor's highest overall index
The fastest
DeepSeek-V4.1-Flash
177 tokens/s, a typical model manages 98
The 20 best language models, with the models from DeepSeek highlighted. “All” and the filter show every model in the index.
20 of 95 models
Provider
Weights
Price (blended per 1M tokens)
Released
Anthropic
OpenAI
Meta
Moonshot AI
DeepSeek
xAI
Z.ai
Xiaomi
Alibaba
StepFun
Mistral AI
MiniMax
Thinking Machines Lab
Inception
| Vendor | Models | Model types | Best language model | Top 10 | Open weights | New in 2026 |
|---|---|---|---|---|---|---|
| Anthropic | 33 | 33 language models | Claude Opus 5.5 (Rank 1 to 3) | 4 | 0 | 13 |
| OpenAI | 63 | 46 language models, 4 image models, 2 video models, 9 audio models, 2 specialist models | GPT-6 Astra (Rank 4 to 8) | 2 | 4 | 27 |
| 66 | 43 language models, 4 image models, 6 video models, 11 audio models, 2 specialist models | Gemini 4 Carbon (Rank 1 to 5) | 2 | 19 | 26 | |
| Meta | 32 | 29 language models, 1 image model, 1 video model, 1 specialist model | Muse Wonder Full (Rank 6 to 9) | 2 | 26 | 7 |
| DeepSeek | 23 | 23 language models | DeepSeek V4.1 Flash 1017 (Rank 13 to 20) | 0 | 23 | 9 |
| Moonshot AI | 12 | 12 language models | Kimi K3.1 Pro (Rank 10 to 13) | 0 | 10 | 7 |
| xAI | 16 | 12 language models, 2 image models, 2 video models | Grok 4.7 (Rank 18 to 24) | 0 | 1 | 12 |
By release date. Grey: retired models.
9 releases since January 2026, by release date.
Prices per 1M tokens for the current language models, according to DeepSeek or the provider linked in each model profile.
| Current model | Input | Output | Context |
|---|---|---|---|
| DeepSeek V4.1 Flash 1017 | $0.15 | $0.60 | 1.05M |
| DeepSeek-V4.1-Flash | $0.15 | $0.60 | 1M |
| DeepSeek-V4-Pro | $0.66 | $1.98 | 1M |
| DeepSeek-V3.2 | $0.28 | $0.42 | 128K |
| DeepSeek-R1 | $0.70 | $2.50 | 128K |
| DeepSeek-V3 0324 | $0.25 | $1 | 164K |
| DeepSeek-V3 | $0.26 | $0.71 | 128K |
| DeepSeek-V4-FlashLimited access | $0.14 | $0.28 | 1M |
| DeepSeek-V3.1 Terminus | $0.27 | $1 | 164K |
| DeepSeek-V3.1 | $0.27 | $1 | 128K |
| DeepSeek-R1 0528 | $0.50 | $2.18 | 164K |
The directory lists 23 DeepSeek models: 23 language models. It counts every model DeepSeek ever released, including retired ones; 22 of them have their own profile. 21 are currently available, the newest being DeepSeek V4.1 Flash 1017, DeepSeek-V4.1-Flash and DeepSeek V4 Flash Vision Exp.
DeepSeek V4.1 Flash 1017. It is at rank 13 to 20 of 95 in the overall index, which combines independent benchmarks from several sources.
DeepSeek V4.1 Flash 1017, released on Oct 17, 2026. The timeline below shows every release by date.
Through the API per 1M tokens (input / output): DeepSeek-V4-Flash $0.14 / $0.28 and DeepSeek V4.1 Flash 1017 $0.15 / $0.60. The prices of all 11 current language models are in the price table.
In the overall index, no. The best OpenAI model, GPT-6 Astra (rank 4 to 8 of 95), is ahead of DeepSeek V4.1 Flash 1017 (rank 13 to 20 of 95). Single tasks can differ.
Yes, 23 of the 23 models have open weights. Each model profile lists how much memory it needs.
DeepSeek V4.1 Flash 1017 with 178 tokens per second according to Artificial Analysis. A typical model manages 98.
The largest is DeepSeek V4.1 Flash 1017 with 1.05 M tokens (≈ 1,570 A4 pages). Every model's value is in the model list.
Data as of Nov 1, 2026