@refuselessai

Abliterated models available via OpenAI-compatible API. No KYC and no gated access.

 United States
refuseless.com
Joined October 2026
  • Tweets97
  • Following31
  • Followers251
7Media
Refuseless — The privacy hub for unrestricted AI. Serving GLM 5.3 Abliterated & Qwen3.8 Abliterated • Zero retention. No logs. • Crypto & XMR accepted
8
3,752
DeepSeek V4.1 Flash 1017 is now FREE on Refuseless.com 🎉 Once the free allowance runs out, you still get below-market rates. Heads up: this is the safeguarded version. The abliterated one is coming soon.
8
12
1
255
11,841
Chinese models ranked by cyber and pentesting capabilities, from personal experience. GLM has done something special
1
3
6
683
Even more underrated is Qwen3.8 Flash Next abliterated
Ahmad
@TheAhmadOsmanOct 15
Qwen 3.8 Flash Next is so underrated
1
3
392
Introducing Qwen3.8 Flash Next Abliterated. 125B total params, 6B active per token. Strong at agentic and cyber tasks. It runs on 260K context window and is the cheapest model we serve, here's the price per million tokens👇 $0.18 in / $0.09 cache / $0.60 out
🤖 Made with AI
1
4,394
New capacity has been added, expect much faster speeds especially on GLM 5.3 Flash!
4
244
GLM 5.3 is still unmatched
Nishant Nischal
@Nishant_N_25Oct 10
Be honest, which is the best open source AI model? 1. DeepSeek 2. Kimi 3. GLM-5 4. Qwen 3.5
2
2
330
Life after you got that huge bug bounty
NOVA
@Its_Nova1012Oct 11
Best career pivot I’ve ever seen.
2
213
500K context is the sweet spot, but I’d let users choose
Tibo
@thsottiauxOct 10
Replying to @rohit3a
Folks. Why do you think we picked 300k for context length for Codex. We ran the numbers and it’s better. We can run it at 1M too but it would just cost more of the usage. Good defaults are important.
1
240
When the weights of Qwen4 will be dropped, we will host an abliterated version and serve it for FREE Follow us to avoid missing out
1
1
9
252
Couldn't agree more.
Wazz
@WazzCryptoOct 9
Some of you aren't using abliterated models to aggressively pentest your vibecoded slop and it shows
2
184
The same prompt tested on Claude Opus 5.5 was immediately blocked by Safeguards.
Refuseless
@refuselessaiOct 9
Here's GLM 5.3 abliterated explaining the network protections that modern Linux kernels are equipped with and how/if they can be bypassed. If you are into cybersecurity, this model is a game changer. We offer it via API on Refuseless.com with zero prompt retention.
1
257
Here's GLM 5.3 abliterated explaining the network protections that modern Linux kernels are equipped with and how/if they can be bypassed. If you are into cybersecurity, this model is a game changer. We offer it via API on Refuseless.com with zero prompt retention.
1
1
1
2
5,748
Who needs an API for Mimo V2.6 Pro Abliterated? Big news coming soon 👀
2
159
Can't wait to try Qwen4 27B abliterated!
2
250
Just use GLM 5.3 abliterated on Refuseless and forget about verification.
0x9527🦀
@coffiasseOct 7
Very efficient🫡 rejected within 5 minutes.
3
250
GLM 5.3 Flash abliterated only costs $0.25/M input and $0.8/M output tokens on Refuseless.com Stronger than Luna and unrestricted!

Refuseless — Hosted API for Abliterated models

OpenAI-compatible hosted inference for a named lineup of refusal-removed open weights.

refuseless.com
Lisan al Gaib
@scaling01Oct 8
the actual highest scores highlighted and consider that Haiku 5.5 is probably much smaller than GLM 5.3 Flash
1
1
216
This is GLM-5.3 abliterated running in opencode, explaining a Linux kernel vulnerability and telling us exactly what the bug is, why it happens, and how to replicate it. No refusal. No lecture. Zero prompt retention. This is the point of Refuseless.
3
20
5,505
Load more