What GPU do you need to run StepFun Step 3.7 Flash?
A new Framework Desktop ($6,799). No priced GPU card holds StepFun Step 3.7 Flash (about 123.6 GB at UD-Q4_K_M); this system holds it with room for its full 256K context, at about 29 tok/s estimate. 2 of the 157 GPUs and systems we track fit it.
201.37B parameters, 11B active (mixture-of-experts) (Hugging Face, accessed 2026-10-18). UD-Q4_K_M GGUF: 122.09 GB (source). Architecture config (Hugging Face config.json; KV heads from num_attention_groups (8); the 3 extra layer_types entries are MTP layers, not counted): 12 global attention layers × 8 KV heads × 128 head dim, plus 33 sliding-window layers (512-token window), so about 48 KB of FP16 KV cache per token of context (how we estimate context).
Card: 198B (196B language + 1.8B vision), about 11B active. Stock llama.cpp support merged Jun 2 (PR #23845). No plain Q4_K_M found (StepFun's own GGUF has Q4_K_S 111.50 GB); the size is unsloth UD-Q4_K_M. Fit ignores the vision projector. Release date is the Hugging Face repo creation date.
Other places to get it
Checked Nov 14, 2026. We list only copies we opened and checked. Check file hashes for copies from other uploaders.
- ModelScope: Other uploaderSame account name as on Hugging Face; we found no link from the lab to it
Single GPUs that fit, cheapest first
| Single GPU that fits | VRAM | Speed estimate | Max context estimate | Price |
|---|---|---|---|---|
| Framework Desktop, AMD Ryzen AI Max+ PRO 495 (Strix Halo), 192 GB | 192 GB | ~29 tok/s | 256K (full) | low $6,799 |
| NVIDIA DGX Spark (128GB) | 128 GB | ~34 tok/s | 94K · 180K (8-bit) | low $6,949.99 |
Every device that runs it
Graphics cards, Macs and mini PCs from the wider hardware list behind My rig, smallest usable memory first. Not sure about yours? Check your exact rig.
Show all 10 devices that run it
| Device | Type | Memory | Speed estimate | Max context estimate | Price |
|---|---|---|---|---|---|
| AMD Ryzen AI Max 385 (128GB) | Mini PC | 128 GB | ~28 tok/s | 94K · 180K (8-bit) | No price yet |
| AMD Ryzen AI Max 390 (128GB) | Mini PC | 128 GB | ~28 tok/s | 94K · 180K (8-bit) | No price yet |
| AMD Ryzen AI Max+ 395 (128GB) | Mini PC | 128 GB | ~28 tok/s | 94K · 180K (8-bit) | low $3,499.99 |
| NVIDIA DGX Spark (128GB) | Mini PC | 128 GB | ~34 tok/s | 94K · 180K (8-bit) | low $6,949.99 |
| NVIDIA RTX Spark N1X laptop (6144-core GPU, 128GB) | Laptop | 128 GB | ~37 tok/s | 94K · 180K (8-bit) | No price yet |
| Apple M2 Ultra 192GB | Mac | 192 GB (~144 usable) | ~69 tok/s | 256K (full) | No price yet |
| Apple M3 Ultra 256GB | Mac | 256 GB (~192 usable) | ~70 tok/s | 256K (full) | No price yet |
| Apple M5 Ultra 256GB | Mac | 256 GB (~192 usable) | ~88 tok/s | 256K (full) | No price yet |
| Framework Desktop, AMD Ryzen AI Max+ PRO 495 (Strix Halo), 192 GB | Mini PC | 192 GB | ~29 tok/s | 256K (full) | low $6,799 |
| Apple M5 Ultra 512GB | Mac | 512 GB (~384 usable) | ~88 tok/s | 256K (full) | No price yet |
Multi-GPU: cards that need two or more
| Cards | Total VRAM | Speed estimate | Max context estimate |
|---|---|---|---|
| 4 × NVIDIA GeForce RTX 5090 | 128 GB | ~169 tok/s | 94K · 180K (8-bit) |
| 3 × NVIDIA RTX A6000 | 144 GB | ~86 tok/s | 256K (full) |
Speed calculator
estimateEstimate = bandwidth ÷ GB read per token (the model's listed GGUF file: Q4_K_M for most models, otherwise the quant named in the list; scaled by active/total parameters for MoE), scaled by a per-platform efficiency and fixed per-token costs calibrated to 45 measured llama.cpp benchmarks (2025–26). Fit = file size + 1.5 GB headroom ≤ total VRAM (on Macs, the part of unified memory macOS lets the GPU use by default), and room for at least 4K tokens of context with an 8-bit KV cache. For a model with no KV cache data, only the file and headroom are checked, so a fit says nothing about context. Multi-GPU: VRAM adds, speed stays one card's (a layer split runs the cards in turn); max context uses the total. Speed is for short context. Max context = what is left after the file and a 1 GB runtime reserve ÷ KV cache bytes per token, default FP16 and 8-bit (q8_0) cache. How speed works · how context works