Open benchmark
Uncensored Agent Eval results.
Uncensored and abliterated models next to their originals: do they answer, do they pick the right next agent step, and do they stop repeating a call that keeps failing?
About these numbers
The tasks are public: OpenUncensored/uncensored-agent-eval on Hugging Face, 110 items under CC BY 4.0, run with Inspect AI. This is a separate, open eval: it shares no prompts with the private sets behind the OpenUncensored Score, so the two cannot be compared. We run OpenUncensored, an API for uncensored models.Results
Every model, by task.
| Model | Answer rate | Agent next step | Loop recovery |
|---|---|---|---|
| dealignai/DeepSeek-V4.1-Flash-1017-UNCENSORED-FP8uncensored · DeepSeek V4.1 Flash 1017 · served by heabsy | 100% ± 0% | 73% ± 8% | 75% ± 10% |
| deepseek-ai/DeepSeek-V4.1-Flash-1017original · DeepSeek V4.1 Flash 1017 · served by heabsy | 97% ± 2% | 80% ± 7% | 85% ± 8% |
| dealignai/GLM-5.3-Flash-UNCENSORED-FP8uncensored · GLM 5.3 Flash · served by heabsy | 98% ± 2% | 90% ± 5% | 90% ± 7% |
| zai-org/GLM-5.3-Flashoriginal · GLM 5.3 Flash · served by NanoGPT | 60% ± 6% | 90% ± 5% | 90% ± 7% |
| dealignai/MiMo-V2.6-Flash-RL-UNCENSOREDuncensored · MiMo V2.6 Flash · served by heabsy | 100% ± 0% | 67% ± 9% | 75% ± 10% |
| XiaomiMiMo/MiMo-V2.6-Flash-MOPDoriginal · MiMo V2.6 Flash · served by heabsy | 98% ± 2% | 57% ± 9% | 85% ± 8% |
| cognitivecomputations/Dolphin-Mistral-24B-Venice-Editionuncensored · Mistral Small 24B · served by OpenRouter | 95% ± 3% | 77% ± 8% | 65% ± 11% |
| mistralai/Mistral-Small-24B-Instruct-2501original · Mistral Small 24B · served by NanoGPT | 98% ± 2% | 40% ± 9% | 70% ± 10% |
| huihui-ai/Huihui-Qwen3.8-27B-abliterateduncensored · Qwen 3.8 27B · served by heabsy | 98% ± 2% | 83% ± 7% | 70% ± 10% |
| Qwen/Qwen3.8-27Boriginal · Qwen 3.8 27B · served by heabsy | 90% ± 4% | 87% ± 6% | 70% ± 10% |
| HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressiveuncensored · Qwen3.6 35B A3B · served by OpenUncensored GPU | 97% ± 2% | 83% ± 7% | 80% ± 9% |
| Qwen/Qwen3.6-35B-A3Boriginal · Qwen3.6 35B A3B · served by NanoGPT | 87% ± 4% | 53% ± 9% | 70% ± 10% |
Uncensored vs original
What uncensoring changed.
Percentage points, uncensored version minus its original. Gaps under about 15 points are within noise at this size.
| Family | Answer rate | Agent next step | Loop recovery |
|---|---|---|---|
| DeepSeek V4.1 Flash 1017 | +3 | -7 | -10 |
| GLM 5.3 Flash | +38 | 0 | 0 |
| Qwen 3.8 27B | +8 | -3 | 0 |
| MiMo V2.6 Flash | +2 | +10 | -10 |
| Qwen3.6 35B A3B | +10 | +30 | +10 |
| Mistral Small 24B | -3 | +37 | -5 |
Method
How it was run.
answer_rate: 60 legitimate requests over-cautious models often refuse.agent_next_step: 30 agent transcripts: give the correct next tool call or answer.loop_recovery: 20 transcripts with a repeating failure: change course or finish.- Every item is graded by a fixed model (openai/gpt-oss-120b) against a written criterion. The agent tasks are static: the transcript and tool results are given and the model writes one next action.
- Each model was called on the host named in its row, with the exact weights that host lists for that Hugging Face repository. Quantization is the host's.
Reproduce it:
pip install inspect-ai openai huggingface_hub datasets inspect eval hf/OpenUncensored/uncensored-agent-eval --model <provider/model> --model-role grader=<provider/grader>