Model comparison
Kadabra Gemma4-E2B SFTJv2-SFT2 vs Claude 4 Sonnet
Capability radar, value bars, and side-by-side posture.
Kadabra Gemma4-E2B SFTJv2-SFT2
Not scored
value score
Claude 4 Sonnet
91
value score
Kadabra Gemma4-E2B SFTJv2-SFT2 ctx
—
Claude 4 Sonnet ctx
200K
Capability radar
Value · context · multimodal · openness · speed posture
Value head-to-head
- Claude 4 Sonnet91
Attribute tape
| Field | Kadabra Gemma4-E2B SFTJv2-SFT2 | Claude 4 Sonnet |
|---|---|---|
| Developer | Kadabra | Anthropic |
| Context | — | 200K |
| Modalities | Text, Code, Image | |
| Openness | — | Proprietary Model |
| Speed | — | Fast |
| Price | — | $3/1M in · $15/1M out |
| Value score | — | 91 |
| Summary | A community fine-tune of Google DeepMind's Gemma 4 E2B open-weight multimodal model, published by Kadabra on Hugging Face with GGUF packaging for efficient local inference. It is an instruction-tuned checkpoint oriented toward resource-constrained deployment rather than a hosted AI service. | Balanced Claude model for coding and agents. |