Home / Alibaba / Qwen / Qwen 3.8 27B
Qwen 3.8 27B API price
qwen/qwen3.8-27b
The cheapest Qwen 3.8 27B API is $0.10/M input at RunInfra (inference host) - all 24 providers compared below, per 1M tokens.
SmallAlibaba / Qwen1M context24 sellers
Input
$0.10
per 1M tokens · RunInfra (inference host)
up to $0.99
Output
$0.40
per 1M tokens · RunInfra (inference host)
up to $4.70
Where to buy it
Cheapest Qwen 3.8 27B API providers compared - 24 sellers, from $0.10/M.
Cost calculator
| Provider | Input $/M | Output $/M | Cached in | TPS | Uptime | Your cost |
|---|---|---|---|---|---|---|
| RunInfra (inference host) | 0.100cheapest | 0.400 | 0.010 | 140 | - | |
| Near AI (via OpenRouter)fp8 | 0.040 | 1.35 | 0.018 | - | 100% | |
| Tenstorrent (via OpenRouter)fp4 | 0.040 | 1.35 | 0.018 | - | 100% | |
| DeepInfra (via OpenRouter)bf16 | 0.150 | 1.88 | 0.037 | - | 100% | |
| AkashML (via OpenRouter)fp8 | 0.225 | 1.98 | 0.050 | - | 100% | |
| Darkbloom (via OpenRouter)fp4 | 0.050 | 2.20 | 0.025 | - | 100% | |
| Cheaper Inference (discount marketplace)35% off | 0.325 | 1.95 | 0.033 | - | - | |
| Wafer (via OpenRouter) | 0.040 | 2.30 | 0.020 | - | 100% | |
| Ionstream (via OpenRouter)fp8 | 0.089 | 2.35 | 0.085 | - | 100% | |
| Chutesfp8 | 0.240 | 2.20 | - | - | 100% | |
| Parasail (via OpenRouter)fp8 | 0.240 | 2.20 | 0.050 | - | 100% | |
| Cerebras (via OpenRouter)fp16 | 0.990 | 1.49 | 0.990 | - | 100% | |
| DekaLLM (via OpenRouter) | 0.040 | 2.50 | 0.020 | - | 100% | |
| DeepInfrabf16 | 0.200 | 2.50 | 0.050 | - | 100% | |
| Mancer 2 (via OpenRouter)fp8 | 0.200 | 2.50 | - | - | 100% | |
| Phala | 0.200 | 2.50 | 0.050 | - | 100% | |
| OrcaRouter (aggregator) | 0.330 | 2.40 | - | - | - | |
| Alibaba (via OpenRouter) | 0.425 | 2.55 | 0.085 | - | 100% | |
| Reka (via OpenRouter)fp8 | 0.040 | 3.00 | 0.020 | - | 100% | |
| CoreWeave (via OpenRouter)fp8 | 0.400 | 3.00 | 0.150 | - | 100% | |
| Novita | 0.420 | 3.00 | 0.085 | - | 100% | |
| Cloudflare (via OpenRouter) | 0.450 | 3.20 | 0.050 | - | - | |
| Venicefp8 | 0.450 | 3.20 | - | - | 100% | |
| ModelRun (via OpenRouter)fp4 | 0.700 | 4.70 | 0.135 | - | 100% | |
| InferX (inference host)promo, 96% off | 0.018 | 0.128 | - | - | - |
24.8× between the cheapest and the dearest of 24 sellers.
Against its tier
- This model$0.50
- Cheapest · Nemotron 3.5 Lightning$0.145
- Tier median · Qwen 3.8 Omni Flash$0.62
- Dearest · GPT 5.4 Mini (03-17 snapshot)$5.25
#8 of 19 small · cheaper than 61%
Details
- Context window
- 1M1,000,000 tokens
- Developer
- Alibaba / Qwen
- Intelligence
- 33.7 Xhigh
- Sellers
- 2424.8× spread
- Cached input
- $0.01where the source publishes one
- Cheapest at
- RunInfra (inference host)
- Combined
- $0.50in + out per 1M
- Last checked
- 2026-10-12
- Throughput
- 140 tok/sas published by RunInfra (inference host)
- Availability
- 100.00%median of 19 OpenRouter endpoints, last 30 min