Qwen3.6 35B Q8 VRAM Requirement and Ollama Model Size
Qwen3.6 35B Q8 needs 44GB minimum VRAM, 54GB optimal VRAM, and 128GB+ system RAM for a safer Ollama run. Popular Ollama model family: Qwen3.6. Caveat: Estimated values are placeholders unless marked measured..
Hardware Snapshot
| Family | Qwen3.6 |
|---|---|
| Scenario | reasoning |
| License scope | open-source |
| Quantization | Q8 |
| VRAM minimum | 44GB |
| VRAM optimal | 54GB |
| Best local GPU | Dual RTX 4090 (model parallel) |
| Cloud fallback | A100 80GB |
| Updated | 2026-02-24 |
| Data status | Verified by Real Hardware |
| Ollama source | Library reference (verified: 2026-02-24) |
| Ollama tag | qwen3.6:35b |
| Category | reasoning |
Benchmark Anchors
| Hardware | Expected tok/s |
|---|---|
| RTX 3090 24GB | 4.9 |
| RTX 4090 24GB | 6.6 |
| A100 80GB | 11.8 |
Quick Answers
- How much VRAM does Qwen3.6 35B Q8 need?
- Qwen3.6 35B Q8 needs about 44GB minimum VRAM and 54GB optimal VRAM for a safer local run target.
- Can Qwen3.6 35B Q8 run on an RTX 3090 24GB?
- Not as a comfortable local target. Qwen3.6 35B Q8 needs about 44GB minimum and 54GB optimal VRAM, so cloud or larger local GPUs are safer.
- What is the Ollama command for Qwen3.6 35B Q8?
- Use ollama run qwen3.6:35b. Check the Ollama tag qwen3.6:35b before running.
- What is the approximate Ollama model size for Qwen3.6 35B Q8?
- Qwen3.6 35B Q8 uses the qwen3.6:35b Ollama tag. Treat the Q8 35 profile as smaller than the runtime VRAM budget; use 44GB minimum VRAM and 54GB optimal VRAM as the safer sizing numbers.
- How much system RAM does Qwen3.6 35B Q8 need in Ollama?
- Plan for at least 128GB system RAM alongside 44GB minimum VRAM and 54GB optimal VRAM. More RAM helps if layers spill to CPU or if you run multiple models.
- What is the RAM vs VRAM requirement for Qwen3.6 35B Q8?
- Qwen3.6 35B Q8 should be sized by GPU VRAM first: 44GB minimum and 54GB optimal. System RAM should be at least 128GB so Ollama has room for CPU-side layers, context, and the surrounding app process.
- What is Qwen3.6 35B Q8 VRAM usage on an RTX 3090?
- Qwen3.6 35B Q8 is not recommended for an RTX 3090 24GB target. The page estimates 44GB minimum and 54GB optimal VRAM, with spill risk marked as high without a larger GPU or cloud.
- Which LocalVRAM profiles share the qwen3.6:35b Ollama tag?
- LocalVRAM tracks these profiles for the same Ollama tag or model family: Qwen3.6 35B Q8 (Q8, 44GB min/54GB optimal); Qwen3.6 35B Q4 (Q4, 22GB min/24GB optimal); Qwen3.6 35B Q5 (Q5, 24GB min/26GB optimal); Qwen3.6 35B FP16 (FP16, 50GB min/62GB optimal). Use the table on this page to compare quantization and VRAM targets before choosing a run command.
- Is Qwen3.6 35B Q8 benchmark data measured or estimated?
- Qwen3.6 35B Q8 has a verified RTX 3090 result of 18.697 tokens/s from LocalVRAM benchmark data.
Ollama Model Size and RAM Notes
| Ollama tag | qwen3.6:35b |
|---|---|
| Run command | ollama run qwen3.6:35b |
| Model size guidance | Q8 quantization, 35 profile; size runtime by VRAM, not download size alone |
| Runtime VRAM target | 44GB minimum, 54GB optimal |
| System RAM planning | 128GB or more when running this tag locally with Ollama |
Download size alone is not a safe runtime budget. Runtime memory also includes KV cache, context length, and GPU layer placement.
Ollama RAM vs VRAM Requirements
| GPU VRAM usage estimate | 44GB minimum / 54GB optimal |
|---|---|
| System RAM requirement | 128GB or more for Ollama runtime headroom |
| RTX 3090 24GB fit | Not recommended |
| CPU spill risk | High without larger GPU or cloud |
| Context warning | Longer context increases KV cache memory, so real VRAM usage can exceed short-prompt estimates. |
For Ollama, VRAM is the first fit check. System RAM matters when layers spill to CPU, when context grows, or when the local app stack runs beside the model.
Available Tag Variants and VRAM Targets
| Profile | Quantization | VRAM target | Ollama tag |
|---|---|---|---|
| Qwen3.6 35B Q8 | Q8 | 44GB min / 54GB optimal | qwen3.6:35b |
| Qwen3.6 35B Q4 | Q4 | 22GB min / 24GB optimal | qwen3.6:35b |
| Qwen3.6 35B Q5 | Q5 | 24GB min / 26GB optimal | qwen3.6:35b |
| Qwen3.6 35B FP16 | FP16 | 50GB min / 62GB optimal | qwen3.6:35b |
Ollama library tags often group multiple quantization choices under one model family. LocalVRAM separates them into VRAM profiles so you can pick the right local target.
Real Hardware Benchmark (RTX 3090)
| Tokens/s | 18.697 |
|---|---|
| Latency | 5915 ms |
| Prompt tokens | 31 |
| Eval tokens | 96 |
| Test time | 2026-08-19T03:08:42Z |
| GPU model | NVIDIA GeForce RTX 3090 |
Verified by real hardware.
Performance Curve
Reference anchors are baseline estimates. Measured RTX 3090 data is overlaid when available.
Best Hardware for Qwen3.6 35B Q8
- Local run: RTX 3090 (24GB) (Check latest deal) for around 18.697 tok/s on this profile.
- Cloud run: RunPod A100 80GB , about 0.6x the local 3090 speed anchor.
- Alternative cloud: Vast.ai options for flexible spot pricing.
Local vs Cloud Cost Hint
| Mode | 40h / month | 120h / month |
|---|---|---|
| Local power only (3090 baseline) | $2.24 | $6.72 |
| A100 80GB | $102 | $306 |
Related Model Profiles
- Qwen3.6 35B Q4 22GB min, 24GB optimal
- Qwen3.6 35B Q5 24GB min, 26GB optimal
- Qwen3.6 35B FP16 50GB min, 62GB optimal
ollama run qwen3.6:35b More reasoning models More 70b-class models Benchmark changelog Submit your test result Run on RunPod Try Vast.ai We may earn a commission if you click links on this page.