MoE premier1.6T total, 49B active per token (MoE). Hybrid attention: Compressed Sparse Attention (CSA, 4x KV compression) + Heavily Compressed Attention (HCA, 128x) across 61 layers, manifold-constrained Hyper-Connections (mHC). FP4 experts + FP8 rest, 32T+ pretraining tokens, Muon optimizer. 1M native context; Think Max mode recommends >=384K. At ~800GB+ for Q4 weights it needs a multi-H200/H100 cluster or DGX-class system - not workstation-fittable, so cloud-only for nearly everyone despite open weights. Top open-weight agentic coding score (SWE-bench Verified 80.6). Open weights under MIT.
- 1600.0B
- 1000k
- mit
- Apr 2026
Scores #
Or run it in the cloud #
Live per-provider pricing, throughput and uptime - refreshed 3 days ago via OpenRouter. Click a column to sort.
| Provider | Type | Input $/M | Output $/M | Cache $/M | Tok/s | Latency | Uptime | Value |
|---|---|---|---|---|---|---|---|---|
| API | 0.44 | 0.87 | - | - | - | - | cheapest |
Default order: throughput among 95%+ uptime providers, then latency; subscriptions last. Sort by any column. Subscription rows show $/mo in the Value column - per-token columns are "-". Affiliate links are marked sponsored / nofollow. Confirm current pricing on the provider's site before committing.
Detailed API pricing page + JSON endpoint →
Inference cost over time #
Data accumulates from the first daily sync - longer ranges populate over time. Prices come from OpenRouter snapshots, not a historical API.