ZAYA1-8B
ZAYA1-8B is a released classification model with open-source; evaluate it while provider pricing coverage matures.
- Family
- ZAYA1
- Released
- 2026-05-06
- Context
- 32K
- Parameters
- 8.4B
- Architecture
- Mixture of Experts
- Specialization
- reasoning
- Openness
- Open source
- License
- Apache 2.0OSI-approvedPermitted
- Weights
- Available
- Code
- Unknown
No tracked provider pricing is available yet.
About
ZAYA1-8B is a small mixture-of-experts (MoE) reasoning model from Zyphra with 8.4B total parameters and approximately 760M active parameters per token. Trained entirely on AMD Instinct MI300X hardware (1,024 GPUs), it incorporates three architectural innovations: Compressed Convolutional Attention (CCA), a novel MLP-based expert router, and learned residual scaling. Post-training uses a four-stage reinforcement learning cascade covering reasoning warmup, adaptive curriculum, large-scale math/code RL with test-time compute traces, and behavioral RL for instruction following. On extended compute (Markovian RSA), it approaches or exceeds frontier models such as DeepSeek-V3.2 and GPT-OSS-120B on mathematics benchmarks.
Providers
No active provider routes are tracked for this model yet.
Capabilities
Benchmark peer barsfor Classification
Benchmark scores(4)
| Benchmark | Score | Version | Evaluation | Source |
|---|---|---|---|---|
| 89.1 | AIME 2026 (accuracy)Observed 2026-10-09 | — | ||
| 71.6 | HMMT Feb 2026Observed 2026-10-09 | — | ||
| 71.0 | diamondObserved 2026-10-09 | — | ||
| 74.2 | MMLU-Pro (accuracy)Observed 2026-10-09 | — |
Migration checks
No linked migration route is available for this model yet.
Rankings & picks(1)
No tracked provider pricing is available yet.