Zarya-4B Release - A100 BF16 - humaneval - pass@1 - create_test: leaderboard
Zarya-4B Release - A100 BF16 - humaneval - pass@1 - create_test — Zarya-4B official native lm-eval results;A100, `dtype=bfloat16`, `attn_implementation=sdpa`, `apply_chat_template`, `slotted_generation=true,slot_size=16,serial_num_blocks=4,slot_threshold=0.9,token_threshold=0.4`. lm-eval0.4.13; native task version,n-shot,filter,metric,stderr retained. Hardware/dtype are separate native conditions,not separate trained models. Slotted speculative decoding,not plain masked-diffusion sampling. Template and exact decoding configuration remain explicit; source reports hardware/configuration sensitivity. Singleton observations remain native; no invented peers, scores or fit. Metric: pass@1 (fraction).
Metric: pass@1 (fraction). Source: huggingface.co. 1 model tracked. Updated 15 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Zarya-4B | 0.0183 | #4590 |
Interactive version: theaggregate.ai/benchmark?slug=zarya-4b-release-a100-bf16-humaneval-pass-1-create-test · How It Works · Data refreshed daily, snapshot 2026-11-16.