Qwen3.8 27B pricing and routing profile
Qwen3.8 27B from Qwen (Alibaba) lists at $0.42 per million input tokens and $3.00 per million output tokens, with a 1,000,000-token context window. Prices from OpenRouter's catalogue on 2026-09-28.
Tool callingStructured outputReasoning controlLog probabilities
Cost per 1,000 requests
| Request shape | List cost |
|---|---|
| Classification or routing (600 tokens in, 10 out) | $0.282 |
| Chat reply (2,000 in, 400 out) | $2.04 |
| Retrieval-augmented answer (8,000 in, 500 out) | $4.86 |
Before prompt caching. Try your own shape in the cost calculator.
Can Gimbix route to it first?
Yes. 12 of 16 tracked hosts return token log probabilities for Qwen3.8 27B, which is the confidence signal Gimbix uses to let a cheaper model answer first.
Measured by Gimbix
On 400 real BANKING77 customer messages, Qwen3.8 27B answered 264 correctly (95% interval 61.2%–70.5%), with 82 invalid answers, at $0.16 per 1,000 requests and a median of 1,634 ms. See all 22 models.
Who serves Qwen3.8 27B
| Host | Input / output per 1M | Log probabilities | Uptime, last 30 min |
|---|---|---|---|
| Darkbloom | $0.05 / $2.20 | yes | 100.0% |
| Wafer | $0.0799 / $4.40 | yes | 100.0% |
| DekaLLM | $0.08 / $2.50 | yes | 99.9% |
| Reka | $0.09 / $4.40 | yes | 100.0% |
| Ionstream | $0.105 / $2.55 | yes | 99.5% |
| DeepInfra | $0.15 / $1.88 | no | 99.7% |
| Phala | $0.199 / $2.08 | no | 90.5% |
| Mancer 2 | $0.20 / $2.50 | yes | 100.0% |
| Parasail | $0.24 / $2.20 | yes | 100.0% |
| Chutes | $0.24 / $2.20 | no | 99.6% |
| AkashML | $0.25 / $2.20 | yes | 99.3% |
| CoreWeave | $0.40 / $3.00 | yes | 100.0% |
| Novita | $0.42 / $3.00 | yes | 99.4% |
| Alibaba | $0.425 / $2.55 | yes | 99.8% |
| Cloudflare | $0.45 / $3.20 | yes | — |
| Venice | $0.45 / $3.20 | no | 99.7% |
A host that lists log probabilities may still omit them; Gimbix checks each one before relying on it.
Cheaper models that can answer first
Models priced below Qwen3.8 27B that expose a confidence signal. Whether they answer your traffic as well is what a Gimbix pilot measures.
| Model | Input / output per 1M | Cheaper on a chat reply |
|---|---|---|
| Qwen3.6 Plus | $0.325 / $1.95 | 30% |
| Qwen3 235B A22B Thinking 2507 | $0.23 / $2.30 | 32% |
| Qwen3 VL 30B A3B Thinking | $0.20 / $2.40 | 33% |
| Qwen3.5-122B-A10B | $0.26 / $2.08 | 34% |
| Qwen3.5 Plus 2026-04-20 | $0.30 / $1.80 | 35% |
| Qwen3 VL 8B Thinking | $0.18 / $2.10 | 41% |
Compare Qwen3.8 27B
- Qwen3.8 27B vs Claude Opus 5.5
- Qwen3.8 27B vs Claude Sonnet 5
- Qwen3.8 27B vs Claude Fable 5.1
- Qwen3.8 27B vs Claude Opus 5
- Qwen3.8 27B vs GPT-6 Sol
- Qwen3.8 27B vs GPT-6 Astra
- Qwen3.8 27B vs GPT-5.6 Sol
- Qwen3.8 27B vs GPT-5.6 Terra
- Qwen3.8 27B vs GPT-5.5
- Qwen3.8 27B vs GPT-5.4
- Qwen3.8 27B vs Gemini 3.1 Pro Preview
- Qwen3.8 27B vs Gemini 3.8 Flash
- Qwen3.8 27B vs Grok 4.7
- Qwen3.8 27B vs Qwen3.8 Max (0902)
- Qwen3.8 27B vs Kimi K3
- Qwen3.8 27B vs GLM 5.3
- Qwen3.8 27B vs DeepSeek V4 Pro 0423
- Qwen3.8 27B vs Mistral Large 3 2512
- Qwen3.8 27B vs Nova Premier 1.0
- Qwen3.8 27B vs Command A
- Qwen3.8 27B vs GPT-6 Luna
- Qwen3.8 27B vs GPT-5.6 Luna
- Qwen3.8 27B vs GPT-5.4 Mini
- Qwen3.8 27B vs GPT-5.4 Nano
- Qwen3.8 27B vs GPT-4.1 Mini
- Qwen3.8 27B vs gpt-oss-120b
- Qwen3.8 27B vs gpt-oss-20b
- Qwen3.8 27B vs Claude Haiku 4.5
- Qwen3.8 27B vs Gemini 3.5 Flash Lite
- Qwen3.8 27B vs Gemma 4 31B
- Qwen3.8 27B vs DeepSeek V4.1 Flash
- Qwen3.8 27B vs DeepSeek V4 Flash 0731
- Qwen3.8 27B vs GLM 5.3 Flash
- Qwen3.8 27B vs Hy4 preview
- Qwen3.8 27B vs MiMo-V2.6-Flash
- Qwen3.8 27B vs Nemotron 3.5 Lightning
- Qwen3.8 27B vs Nemotron 3 Ultra
- Qwen3.8 27B vs Qwen3.8 Flash
- Qwen3.8 27B vs Qwen3.7 Flash
- Qwen3.8 27B vs MiniMax M3
- Qwen3.8 27B vs Kimi K2.6
- Qwen3.8 27B vs Mistral Small 4
- Qwen3.8 27B vs Ministral 3 8B 2512
- Qwen3.8 27B vs Llama 4 Maverick
- Qwen3.8 27B vs Llama 3.3 70B Instruct
- Qwen3.8 27B vs Nova Lite 1.0
- Qwen3.8 27B vs Nova Micro 1.0
- Qwen3.8 27B vs Command R7B (12-2024)
- Qwen3.8 27B vs Phi 4
Questions
How much does Qwen3.8 27B cost?
Qwen3.8 27B lists at $0.42 per million input tokens and $3.00 per million output tokens on OpenRouter as of 2026-09-28. A typical chat reply of 2,000 input and 400 output tokens costs about $2.04 per 1,000 requests.
What is the context window of Qwen3.8 27B?
1,000,000 tokens.
Can a cheaper model replace Qwen3.8 27B?
Sometimes. Gimbix proves it on your own traffic before routing: the cheaper model must match Qwen3.8 27B on one half of your requests and make at most three answers worse on the other.