Can I run ThinkingCap-Qwen3.8-Flash-Next on an NVIDIA H200?
Probably, at 4-bit
The published weights are too big, but a 4-bit version (about 101 GB) would fit, with room for ≈ 256K tokens of context. The site hasn't analyzed a GGUF of ThinkingCap-Qwen3.8-Flash-Next yet.
Files and context on NVIDIA H200
Each file of ThinkingCap-Qwen3.8-Flash-Next (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.
| File | Size | On NVIDIA H200 | Context (16-bit) | Context (8-bit) |
|---|---|---|---|---|
| As published | 360 GB | Doesn't fit | — | — |
| 4-bit (estimate) | 101 GB | Fits | 256K | 256K |