DeepSeek
Verdict
Evidence: Editorial assessmentDeepSeek publishes capable open-weight models under the permissive MIT licence, which makes them popular for self-hosting. Its GitHub lists DeepSeek-V3 and V3.2-Exp; newer V4-series releases were reported in mid-2026 but had no repository when we checked. Data-residency rules matter if you use DeepSeek's own hosted app or API.
Quick facts
| DeepSeek | |
|---|---|
| Maker | DeepSeek |
| Category | AI Models & Open-Weight LLMs |
| Pricing model | Open weights |
| Status | Active |
| Official site | github.com |
| Data last checked | 5 October 2026 |
Key facts and sources
| Claim | Detail | Evidence | Source |
|---|---|---|---|
| Open-weight repositories | DeepSeek publishes DeepSeek-V3 (MIT licence) and DeepSeek-V3.2-Exp (MIT licence, updated November 2025) on GitHub. No V4 model repository was listed on 5 October 2026. | Verified | deepseek-ai on GitHub, 5 October 2026 |
| Notable releases, July to September 2026 | Reported: Kimi K3 open weights (27 July), DeepSeek V4-Flash (31 July) and V4.1 Flash under MIT licence (10 September), GLM-5.3 under MIT (August), Qwen3.8-Max (API only), and a Gemma 4 update (16 July). | Reported | ThursdAI release log, July 2026, 5 October 2026 |
Pricing
No prices are shown because we have not confirmed them on DeepSeek's own pricing page. Go to DeepSeek
Who should use it
- Self-hosting under the MIT licence
- Cost-sensitive teams evaluating open models
Who should skip it
- Your policies restrict services hosted in China (self-host instead)
Limitations and risks
- V4-series details are reported, not confirmed on a DeepSeek repository.
- API prices were not verified and are not shown.
How it compares: DeepSeek alternatives
Qwen
Alibaba's Qwen family of language models; Qwen3.8 was released as open weights in August 2026, including a 2.4T-parameter flagship.
Kimi
Moonshot AI's Kimi models; Kimi K3 is a 2.8T-parameter open-weight model with native vision and a 1M-token context window.
gpt-oss
OpenAI's open-weight reasoning models, gpt-oss-120b and gpt-oss-20b, released under the Apache 2.0 licence.
Gemma
Google DeepMind's family of open-weight language models; Gemma 3n is the mobile-first member and Gemma 4 is the newest generation.
Llama 2
Meta's 2023 open-weight language model family. It is a legacy release, superseded by newer Llama and other open-weight models.
Rating
Not yet rated. We show a rating only when at least four of six criteria have documented evidence. How we rate AI tools
Common questions
Is DeepSeek open source?
DeepSeek-V3 and DeepSeek-V3.2-Exp are published on GitHub under the MIT licence. Check each model card for weight terms.
What is the newest DeepSeek model?
V4-series models (V4-Flash, V4.1 Flash) were reported in July and September 2026, but no V4 repository was listed on DeepSeek's GitHub on 5 October 2026.
Sources
- deepseek-ai on GitHub · checked 5 October 2026
- ThursdAI release log, July 2026 · checked 5 October 2026
- DeepSeek official site