Size, context and price
Size
Total parameters, billions. Log scale. Larger is more capable
Context window
Tokens the model holds at once. Higher is better
Price
Blended USD per 1M tokens, lowest listed host. Lower is better
AbliteratedFine-tunedHatched: weights only, no API listed
By how refusals were removed
Abliterated
The refusal direction is found in the activations and projected out of the weights.
- #1Kimi K3.1 Pro2,780B
- #2DeepSeek V4.1 Flash 1017763B
- #3GLM-5.3753B
- #4GLM-5.3 Flash320B
- #5Qwen 3.6 35B-A3B Abliterated36B
- #6Gemma 4 31B Heretic31.3B
- #7Qwen 3.8 27B Abliterated27.8B
- #8Llama 3.1 8B Abliterated8B
Size8 models
Fine-tuned
The model is trained again on data that answers instead of refusing.
- #1Hermes 3 405B405B
Size1 model
Every model
| Refusals | Size | Context | Features | Price | Availability | ||
|---|---|---|---|---|---|---|---|
| Model | How refusals were removed | Parameterstotal | Context windowtokens | Inputs | In / outUSD per 1M | Where it runs | Further analysis |
#1Kimi K3.1 Pro Moonshot AI |
AbliteratedWeights edited | 2,780B | 1M | $3.50 / $8.00 | Silk Compute |
Model Compare | |
#2DeepSeek V4.1 Flash 1017 DeepSeek |
AbliteratedWeights edited | 763B | 1M | $0.24 / $0.90 | Silk Compute+2 more |
Model Compare | |
#3GLM-5.3 Z.ai |
AbliteratedWeights edited | 753B | 1M | $2.00 / $4.00 | Silk Compute+1 more |
Model Compare | |
#4GLM-5.3 Flash Z.ai |
AbliteratedWeights edited | 320B18B active | 1M | $0.24 / $1.00 | Silk Compute |
Model Compare | |
#5Hermes 3 405B Nous Research |
Fine-tunedRetrained | 405B | 128K | $1.20 / $1.20 | Unfil.AI |
Model Compare | |
#6Qwen 3.6 35B-A3B Abliterated Qwen |
AbliteratedWeights edited | 36B3B active | 256K | Local only | Your machine24 GB of GPU memory |
Model Compare | |
#7Gemma 4 31B Heretic Google |
AbliteratedWeights edited | 31.3B | 256K | Local only | Your machine24 GB of GPU memory |
Model Compare | |
#8Qwen 3.8 27B Abliterated Qwen |
AbliteratedWeights edited | 27.8B | 256K | $1.60 / $12.00 | Featherlessor locally |
Model Compare | |
#9Llama 3.1 8B Abliterated Meta |
AbliteratedWeights edited | 8B | 128K | Local only | Your machine8 GB of GPU memory |
Model Compare |