Runs on ≤16 GB · Hugging Face ↗
Coding index 20.7 (Artificial Analysis)
OpenAI's open-weight 20B, shipped natively in MXFP4 (~4-bit) — so its on-disk size already IS the quantized size. Uses the "harmony" chat format with separate reasoning channels.
HumanEval pass@1 97% (29/30), via local GPU (Ollama, Q4_K_M). Measured by us on this hardware, as a rough sanity check — HumanEval is a different, easier, partly-contaminated benchmark than Artificial Analysis’ composite, so it is not comparable to the coding-index column.
BigCodeBench-Hard pass@1 26% (31/121), via local GPU (Ollama, MXFP4). The brutal counterpart to HumanEval — where HumanEval saturates near the top, BCB-Hard spreads the field, so this is the number that actually separates coding ability.
~61 tok/s on RTX 4060 Ti, measured by us.
Real GGUF file sizes = weight VRAM. Add the KV cache for your context (≈2 GB at 32K, fp16).
| Quantization | Size |
|---|---|
| Q8_0 | 0.9 GB |
| BF16 | 1.7 GB |
| MXFP4 | 12.1 GB |
Do NOT re-quantize — it is already 4-bit; a Q8 re-quant only wastes VRAM. Fits ~13 GB. Needs a runtime with MXFP4 support (recent llama.cpp / vLLM / Ollama). Set the reasoning effort via the harmony system fields.