Index / providers / featherless-ai

Featherless.ai

Serverless host for Hugging Face open-weight models. The largest catalog of community abliterated, Heretic, Dolphin, and Hermes checkpoints we found, behind an OpenAI-compatible API.

Models25

Per-token price$0.058–$4.20/1M tok

Plans from$25/mo

Free tier—

featherless.ai ↗Pending review

[ 01 ] Models & prices

BILLINGChat plan is a flat $25/month with unlimited tokens, 32K context and 4 concurrency units (each model costs 1-4 units by size); Developer plan is a monthly credit amount from $50/month billed per token at each model's input/output rate (only successful requests bill, credits don't expire). Paid by credit/debit card via Stripe; Business plans can be invoiced.

Model	Method	Context	In $/1M	Out $/1M	Per 1K chats
mlabonne/Meta-Llama-3.1-8B-Instruct-abliteratedLlama-3.1-8B-Instruct · 8B · Llama 3.1 Community	Abliterated	33K	$0.050	$0.080	$0.090
huihui-ai/Huihui-gemma-4-31B-it-abliterated-v2gemma-4-31B-it · 31B · Apache-2.0	Abliterated	33K	$0.12	$0.36	$0.30
NousResearch/Hermes-3-Llama-3.1-8BLlama-3.1-8B · 8B · Llama 3.1 Community	Fine-tuned	33K	$0.20	$0.32	$0.36
dphn/Dolphin3.0-Llama3.1-8BLlama-3.1-8B · 8B · Llama 3.1 Community	Fine-tuned	33K	$0.20	$0.32	$0.36
coder3101/gemma-4-26B-A4B-it-hereticgemma-4-26B-A4B-it · 26B · Apache-2.0	Abliterated	33K	$0.40	$0.80	$0.80
NousResearch/Hermes-4-14BQwen3-14B · 14B · Apache-2.0	Fine-tuned	33K	$0.48	$0.96	$0.96
mlabonne/Qwen3-14B-abliteratedQwen3-14B · 14B · Apache-2.0	Abliterated	33K	$0.48	$0.96	$0.96
DavidAU/Qwen3.5-9B-Claude-4.6-HighIQ-INSTRUCT-HERETIC-UNCENSOREDQwen3.5-9B · 9B · Apache-2.0	Abliterated	33K	$0.43	$1.12	$0.99
huihui-ai/Llama-3.3-70B-Instruct-abliteratedLlama-3.3-70B-Instruct · 70B · Llama 3.3 Community	Abliterated	33K	$0.65	$0.75	$1.02
llmfan46/gemma-4-31B-it-uncensored-hereticgemma-4-31B-it · 31B · Apache-2.0	Abliterated	33K	$0.48	$1.44	$1.20
wangzhang/gemma-4-31B-it-abliteratedgemma-4-31B-it · 31B · Apache-2.0	Abliterated	33K	$0.48	$1.44	$1.20
NousResearch/Hermes-3-Llama-3.1-70BLlama-3.1-70B · 70B · Llama 3.1 Community	Fine-tuned	33K	$0.72	$0.72	$1.08
huihui-ai/Qwen3-32B-abliteratedQwen3-32B · 32B · Apache-2.0	Abliterated	33K	$0.41	$1.97	$1.39
huihui-ai/Huihui-Qwen3-8B-abliterated-v2Qwen3-8B · 8B · Apache-2.0	Abliterated	33K	$0.47	$1.82	$1.38
dphn/Dolphin-Mistral-24B-Venice-EditionMistral-Small-24B-Instruct-2501 · 24B · Apache-2.0	Fine-tuned	33K	$0.70	$1.16	$1.28
huihui-ai/Huihui-Qwen3.5-27B-abliteratedQwen3.5-27B · 27B · Apache-2.0	Abliterated	33K	$0.30	$2.40	$1.50
dphn/dolphin-2.9.3-mistral-nemo-12bMistral-Nemo-12B · 12B	Fine-tuned	33K	$0.87	$0.99	$1.36
huihui-ai/Huihui-Mistral-Small-3.2-24B-Instruct-2506-abliteratedMistral-Small-3.2-24B-Instruct-2506 · 24B · Apache-2.0	Abliterated	33K	$0.87	$1.61	$1.67
huihui-ai/Huihui-Qwen3.6-35B-A3B-abliteratedQwen3.6-35B-A3B · 35B	Abliterated	33K	$0.40	$4.00	$2.40
llmfan46/GLM-4-32B-0414-uncensored-heretic-v2GLM-4-32B-0414 · 32B	Abliterated	33K	$1.06	$2.60	$2.36
mlabonne/Llama-3.1-70B-Instruct-lorablatedLlama-3.1-70B-Instruct · 70B · Llama 3.1 Community	Abliterated	33K	$2.88	$2.88	$4.32
NousResearch/Hermes-4-70BLlama-3.1-70B · 70B · Llama 3.1 Community	Fine-tuned	33K	$2.88	$2.88	$4.32
huihui-ai/Huihui-Qwen3.8-27B-abliteratedQwen3.8-27B · 27B	Abliterated	33K	$1.60	$12.00	$7.60
heretic-org/Qwen3.8-27B-heretic-araQwen3.8-27B · 27B	Abliterated	33K	$1.60	$12.00	$7.60
ArliAI/Qwen3.5-27B-DerestrictedQwen3.5-27B · 27B · Apache-2.0	Abliterated	33K	$1.60	$12.00	$7.60

Per 1K chats = cost of 1,000 messages of ~1,000 tokens in and ~500 out. Click a model to compare it across providers.All models →

[ 02 ] At a glance

Method: Mixed
Access: Web chat + OpenAI-compatible API
Logging: No logs
Trains on prompts: Unverified
Guarantee: Policy only
Jurisdiction: Not disclosed
Signup: Account required
Payments: Card
Data checked: Oct 11, 2026

PRIVACYStates it doesn't collect, store, or log prompts or completions sent through the API. Training use and jurisdiction aren't addressed. Policy ↗

[ 03 ] Scores

Compliance

Benchmark pending

measured

Capability

Benchmark pending

measured

Privacy

23.0

from published policy & pricing

Truthfulness

Benchmark pending

measured

Cost

97.7

from published policy & pricing

Speed

Benchmark pending

measured

Composite: pending capability and compliance runs · method

[ 04 ] Notes & sources

The public model API listed 22,182 models on 2026-10-11 (marketing says 40,000+). 277 had "abliterated" in the ID, and about 1,000 matched uncensored, abliterated, Heretic, Dolphin, Hermes, or Derestricted. We list 25 popular picks by downloads; search the full catalog on featherless.ai. Every uncensored model is served at 32K context through the API (checked against the model API on 2026-10-14), even where the base model supports more. The company grew out of the RWKV project. Card payments via Stripe; its legal entity isn't stated on the pages we checked.