Qwen · Chat and reasoning · 235B parameters · 22B active · apache-2.0
Qwen3 235B A22B Thinking 2507
Needs 282 GB at FP8 / INT8. The smallest real configuration today is 4× A100 80GB at Vast.ai, $2.40 per hour, about $1,754 a month at 730 hours.
Parameters235B22B active per token
Context256k tokensmaximum published
Checkpoint470 GBBF16, as published
Licenseapache-2.0check before commercial use
Released—documented date
Smallest configuration · FP8 / INT8
4× A100 80GB320 GB VRAM · 88% used
- Per hour
- $2.40
- Per month
- $1,754
- Cheapest at
Vast.ai
Memory · 3 precisions · 5 context lengthsHow much VRAM it needs
Weights plus runtime overhead, then the KV cache that grows with the context you serve.
| Precision | Weights | Weights + headroom | 4k context | 32k context | 128k context | 256k context |
|---|---|---|---|---|---|---|
| BF16published checkpoint | 470 GB | 565 GB | 566 GB | 571 GB | 590 GB | 615 GB |
| FP8 / INT8estimated | 235 GB | 282 GB | 283 GB | 289 GB | 308 GB | 333 GB |
| INT4 / MXFP4estimated | 141 GB | 170 GB | 170 GB | 176 GB | 195 GB | 220 GB |
| KV cacheone sequence, FP16 cache, included above | — | — | 1 GB | 7 GB | 26 GB | 51 GB |
Cross-checked on Hugging Face Oct 28, 2026: 235.1B parameters in the safetensors headers, 470.2 GB of weight files, attention layout from config.json (94 layers, 4 KV heads × 128). Method: weights as published at the native precision, verified on Hugging Face 2026-10-28; other precisions estimated at 1 (8-bit) and 0.6 (4-bit) bytes per parameter; × 1.2 runtime headroom; KV cache = 2048 bytes per token per layer × 94 full-attention layers (FP16 cache, one sequence).config.json ↗
Smallest configuration per GPU · 109 GPUs · cheapest eligible on-demand offer per GPU count, every provider on fileWhat it costs to run, GPU by GPU
The fewest GPUs of one type whose total memory fits the model at short context, priced from the cheapest eligible on-demand offer for exactly that count (45-day window, newest reading per provider), with the next providers beside it. Monthly = 730 hours.
| GPU | Smallest config | VRAM used | $ / hour | $ / month | Cheapest at |
|---|---|---|---|---|---|
| A100 80GBBest fit | 4× 80 GB320 GB total | 88% | $2.40 | $1,754 | Vast.aiTrust 78 · read Nov 2 · receipt · also Edgevana $4.60, Gcore $5.63 |
| A100 40GB | 8× 40 GB320 GB total | 88% | $3.20 | $2,338 | Vast.aiTrust 78 · read Nov 2 · receipt · also Denvr Cloud $9.20, Sesterce Cloud $12.32 |
| RTX A6000 | 8× 48 GB384 GB total | 73% | $3.36 | $2,453 |  LiumTrust 40 · read Nov 2 · receipt · also LeaderGPU $3.58, Verda $6.62 |
| L40S | 8× 48 GB384 GB total | 73% | $3.74 | $2,727 | Vast.aiTrust 78 · read Nov 2 · receipt · also Lium $6.00, GPU.ai $6.40 |
| RTX PRO 6000 | 4× 96 GB384 GB total | 73% | $3.84 | $2,805 | Vast.aiTrust 78 · read Nov 2 · receipt · also UpCloud $3.89, Hetzner $6.44 |
| MI350X | 1× 288 GB288 GB total | 98% tight | $3.99 | $2,913 |  OpenRelaynot scored · read Nov 2 · receipt |
| Gaudi 2 | 4× 96 GB384 GB total | 73% | $4.18 | $3,052 |  Cyfuture AInot scored · read Nov 2 · receipt |
| A40 | 8× 48 GB384 GB total | 73% | $4.61 | $3,364 |  LeaderGPUTrust 59 · read Nov 2 · receipt · also Exoscale $8.36 |
| A100X | 4× 80 GB320 GB total | 88% | $5.12 | $3,739 | Vast.aiTrust 78 · read Nov 2 · receipt |
| RTX 6000 Ada | 8× 48 GB384 GB total | 73% | $5.20 | $3,796 |  GPU.ainot scored · read Nov 2 · receipt · also io.net $5.46, QuantaCloud $6.21 |
| H200 NVL | 2× 141 GB282 GB total | 100% tight | $5.33 | $3,894 | Vast.aiTrust 78 · read Nov 2 · receipt · also IONOS $6.52, Geodd $6.78 |
| MI300X | 2× 192 GB384 GB total | 73% | $5.71 | $4,167 | Cyfuture AInot scored · read Nov 2 · receipt · also Hot Aisle $5.98, AtmosCompute $6.69 |
| MI355X | 1× 288 GB288 GB total | 98% tight | $5.99 | $4,373 |  Daytonanot scored · read Nov 2 · receipt · also OpenRelay $6.30 |
| MI325X | 2× 256 GB512 GB total | 55% | $6.21 | $4,536 | Cyfuture AInot scored · read Nov 2 · receipt |
| H100 | 4× 80 GB320 GB total | 88% | $6.84 | $4,991 |  GcoreTrust 80 · read Nov 2 · receipt · also GPU.ai $7.92, Edgevana $7.98 |
| H200 | 2× 141 GB282 GB total | 100% tight | $7.38 | $5,387 |  Theta EdgeCloudnot scored · read Nov 2 · receipt · also Vast.ai $7.53, Canopy Wave $8.00 |
| B200 | 2× 180 GB360 GB total | 78% | $8.22 | $6,002 |  UpCloudTrust 75 · read Nov 2 · receipt · also Deep Infra $9.78, Elice AI Cloud $13.64 |
| L40 | 8× 48 GB384 GB total | 73% | $8.40 | $6,132 |  OblivusTrust 82 · read Nov 2 · receipt · also CoreWeave $10.00, AtmosCompute $12.21 |
| B300 | 2× 270 GB540 GB total | 52% | $11.65 | $8,506 | UpCloudTrust 75 · read Nov 2 · receipt · also FPT AI Factory $13.98, Vast.ai $17.00 |
| H100 NVL | 4× 94 GB376 GB total | 75% | $17.86 | $13,038 |  AtmosComputenot scored · read Nov 2 · receipt · also EmpirioLabs AI $22.40, Open Telekom Cloud $25.90 |
| B200 CC | 2× 180 GB360 GB total | 78% | $18.09 | $13,206 |  VerdaTrust 81 · read Nov 2 · receipt |
| GB300 | 2× 279 GB558 GB total | 51% | $26.11 | $19,060 | VerdaTrust 81 · read Nov 2 · receipt |
| B300 CC | 4× 270 GB1080 GB total | 26% | $46.31 | $33,806 | VerdaTrust 81 · read Nov 2 · receipt |
| A800 80GB PCIe | 8× 80 GB640 GB total | 44% | $50.80 | $37,088 |  Volcano Enginenot scored · read Nov 2 · receipt |
Throughput estimates · Oct 4, 2026Tokens per second, and what a million costs
Chat and reasoning. Speed estimated from memory bandwidth and model size, never measured. The cost per 1M output tokens is an estimate at the prices read on Oct 4, 2026; it is not recomputed here.
| GPU | Cards | Precision | tok/s per user | per 1M output tokens · estimate |
|---|---|---|---|---|
| MI300X | 2 | FP8 | 34.3 | $1.08 |
| B200 | 2 | FP8 | 60.5 | $1.44 |
| H100 SXM | 4 | FP8 | 35.2 | $1.88 |
| RTX PRO 6000 | 4 | FP8 | 22.2 | $2.39 |
| H100 PCIe | 4 | FP8 | 21.0 | $2.63 |
| H200 | 2 | FP8 | 44.0 | $3.00 |
| L40S | 8 | FP8 | 15.1 | $3.20 |
| A100 80GB | 8 | BF16 | 22.5 | $3.44 |
cards, precision and speed estimated from memory bandwidth and model size; cost recomputed by the site from verified prices · cheapest per output at the prices of Oct 4, 2026: MI300XWhy estimates, and their limits →
Qwen3 235B A22B Thinking 2507 · related