DeepSeek V4.1 Flash is DeepSeek's fast open weight model: a mixture of experts released under the MIT license, with a context window of up to a million tokens. There are two checkpoints: the original from September 10, 2026 (often labelled 0910) and the October 17 update (1017), which scores higher on hard reasoning. It is also the most used model family on Open Uncensored.
Because the weights are open, several groups have published uncensored versions: abliterated, heretic edited and fine tuned. They are not interchangeable. We run all of them through the same tests, and this page compares every one we serve.
The originals
- DeepSeek V4.1 Flash 1017 scores 43% on Humanity's Last Exam (Artificial Analysis, reasoning mode).
- DeepSeek V4.1 Flash scores 39% on Humanity's Last Exam and 91% on GPQA Diamond (Artificial Analysis, reasoning mode).
Every uncensored version keeps that starting point. What differs is how much each edit costs in agent reliability, how much it still refuses, and the price.
Every uncensored DeepSeek V4.1 Flash on Open Uncensored
| Model | id | Rank (score) | Agent tasks | Loops | Answers edgy requests | Hard-line refusals | $ per 1M in + out | Context | Providers |
|---|---|---|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017 Uncensored | deepseek-v4.1-flash-1017-uncensored |
#1 (69.8) | 2/5 | 0% | 100% | 0% | $1.65 | 1049K | venice, nanogpt |
| DeepSeek V4.1 Flash 0910 Uncensored | deepseek-v4.1-flash-0910-uncensored |
#4 (66.1) | 2/5 | 20% | 97% | 0% | $1.65 | 524K | nanogpt |
| DeepSeek V4.1 Flash Uncensored | deepseek-v4.1-flash-uncensored |
#5 (64.3) | 1/5 | 0% | 100% | 0% | $1.65 | 1049K | venice, nanogpt |
| Abliterated Large Flash | abliterated-large-flash |
#6 (64.0) | 1/5 | 20% | 92% | 0% | $3.00 | 1049K | abliteration-ai, venice, nanogpt |
| DeepSeek V4.1 Flash Heretic | deepseek-v4.1-flash-1017-heretic |
not ranked yet | – | – | 94% | 0% | $0.990 | 512K | nanogpt |
Reasoning variants with the same weights: deepseek-v4.1-flash-0910-uncensored-thinking, deepseek-v4.1-flash-1017-heretic-thinking, deepseek-v4.1-flash-1017-uncensored-thinking, deepseek-v4.1-flash-uncensored-thinking. They think before answering, which helps on hard problems and costs more tokens.
Which one to use
- Best overall: DeepSeek V4.1 Flash 1017 Uncensored (rank #1, score 69.8).
- For agents and coding tools: DeepSeek V4.1 Flash 1017 Uncensored again, with 2 of 5 agent tasks finished and 0% loops.
- Most open: DeepSeek V4.1 Flash 1017 Uncensored, answers 100% of edgy requests and refuses 0% of hard-line ones.
- Cheapest: DeepSeek V4.1 Flash Heretic at $0.990 per million input plus output tokens.
Agent tasks are five multi-step tool scenarios, one run each, so a one task gap is noise. Answers edgy requests counts 36 legitimate prompts that cautious models refuse. The live numbers, including speed on real traffic, are on the rankings page.
Call it
curl https://openuncensored.com/api/v1/chat/completions \
-H "Authorization: Bearer $OU_API_KEY" -H "Content-Type: application/json" \
-d '{"model": "deepseek-v4.1-flash-1017-uncensored", "messages": [{"role": "user", "content": "Hello"}]}'
If the providers for a version are down, the API falls back to the next best ranked uncensored DeepSeek V4.1 Flash built on
the same original, and the X-OU-Model header tells you which one answered.
Try any of them in a minute
No account and no sign up. Buy a prepaid key at openuncensored.com/credits, or hand your agent openuncensored.com/agents.md and let it get one for you. Then:
curl https://openuncensored.com/api/v1/chat/completions \
-H "Authorization: Bearer $OU_API_KEY" -H "Content-Type: application/json" \
-d '{"model": "deepseek-v4.1-flash-1017-uncensored", "messages": [{"role": "user", "content": "Hello"}]}'
The full, always current table is on the rankings page.