Skip to content
Hot eventDeveloping

bf16.ai launches uncensored DeepSeek and GLM APIs

1 reports1 reporting sourcesUpdated 13 h ago

Start here

AI digest

bf16.ai has launched API access to uncensored versions of DeepSeek and GLM models, served through native OpenAI and Anthropic endpoints with a 1M-token context window. The company says refusal behavior was removed through weight-level editing rather than retraining, and that it does not inject a chat filter, though it may block clearly unlawful traffic at its gateway or upstream infrastructure. The lineup includes deepseek-v4.1-flash-uncensored at $0.24/$0.90 per 1M input/output tokens and glm-5.3-uncensored at $1.40/$4.40, which bf16.ai says is at or below official base API rates. Billing is pay-per-token, with $0.50 in free credit valid for 30 days and no card required; unused purchased credit is refundable within 14 days (minimum $1), while free credit is not. The company says it keeps request metadata for 90 days and billing records for 7 years, and that no prompts are logged. The models are positioned for use in Claude Code, Codex, OpenClaw and Hermes Agent.

Generated from reports · updated 13 h ago

Report timeline

Follow the reports to see the event from different sides.

November 16
  1. Hacker News · AI
    bf16.ai Launches Uncensored DeepSeek and GLM APIs with 1M Context and No Prompt Logging

    bf16.ai now serves uncensored DeepSeek and GLM checkpoints via native OpenAI and Anthropic endpoints with a 1M-token context window and no prompt logging. The lineup includes deepseek-v4.1-flash-uncensored at $0.24/$0.90 per 1M input/output tokens and glm-5.3-uncensored at $1.40/$4.40, priced at or below official base API rates with pay-per-token billing and $0.50 free credit. Refusal behavior is removed through weight-level editing rather than retraining, and the models plug into Claude Code, Codex, OpenClaw, and Hermes Agent.

Heat for this event

Current, comparable range
7
Peak, comparable range
10November 16 04:00
Change in 24 hours
–

The trend compares the same participants over hours observed continuously, so its range can be narrower than the current heat count. Hover or tap the chart for hourly heat; the arrow keys move through it.