Qwen 3.8 27B
qwen/qwen3.8-27bqwen3-8-27bqwen3.8-27b- 🧠 27B dense parameters, reasoning and code-optimized
- 📏 262K token context window
- 🔒 Runs inside a Trusted Execution Environment with attestation
- 👁️ Multimodal: accepts text and image input
- 🔧 Function calling and web search supported
- 📜 Apache 2.0 licensed and openly available
Alibaba Group is a Chinese multinational technology company founded in 1999 and headquartered in Hangzhou, Zhejiang. Originally built around e-commerce and cloud computing, Alibaba has become one of the most prolific contributors to open-weight AI research, developing the Qwen…
Explore 51 more models by Alibaba Group →Qwen 3.8 27B in its TEE configuration pairs Alibaba's 27-billion-parameter multimodal text model with hardware-enforced confidential computing, so inference runs inside a sealed enclave and users can independently verify enclave identity and configuration through attestation evidence. Functionally it matches the standard Qwen 3.8 27B deployment released in August 2026 — a 262K-token context window, image plus text input, reasoning, code specialization, function calling and web search — but adds a cryptographic guarantee about where and how the computation happened.
Within Alibaba's Qwen lineup, the 27B sits in the mid-size dense tier: substantially more capable than small models like Qwen 3.5 9B, while far lighter to serve than sparse flagships such as Qwen 3.8 2.4T or Qwen 3.5 397B. Its Apache 2.0 license and very large HuggingFace following reflect how widely the Qwen dense checkpoints are used in production.
It is a strong fit for privacy-sensitive workloads that still need real capability — confidential document analysis over long contexts, agentic tool-calling pipelines, code assistance, and vision-grounded tasks where verifiable execution matters as much as output quality.
This About section is AI-generated from public sources via VeniceStats + Venice inference, with no human editing. It may contain inaccuracies.
Daily usage · Qwen 3.8 27B · Organic sellers
| Seller | Health | Rep↓ | Sales | Tokens | Routing | Input $/M | Cached $/M | Output $/M | Save | 1st token | tok/s |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Open ForgeSAME 0x1d90…b0aa | healthy | 91.11 | $6.93 | 6.52M | #2 | $0.50 | $0.10 | $3.00 | — | 1.1 s | 57 |
| 💬 BanterSAME 0x4122…e194 | healthy | 82.88 | $0.02 | 149.80k | #1 | $0.105 | $0.0261 | $1.31 | −30% | — | — |
| Relay Works 0x3040…7139 | — | 54.66 | $0.11 | 286.35k | gated | $0.14 | $0.035 | $1.75 | −7% | — | — |
| attacompute.ai 0xd5e7…7be0 | healthy | 51.22 | $0.15 | 178.49k | gated | $0.396 | $0.0495 | $2.97 | — | — | — |
| RedPillSAME 0x7e99…dc4f | healthy | 43.30 | $0.83 | 413.22k | gated | $0.20 | $0.05 | $2.50 | — | — | — |
| SCloud 0x0da6…cd0c | healthy | 25.00 | — | — | gated | $0.006 | $0.0006 | $0.03 | −96% | — | — |
| Unified Rack · DeepSeek 240 tok/sSAME 0x42d1…822f | healthy | 13.21 | $0.03 | 81.72k | gated | $0.0469 | $0.0162 | $1.7803 | −69% | — | — |
| OpenCheapSAME 0x0ee1…6c41 | — | 12.43 | $0.01 | 2.70M | gated | $0.0049 | $0.0007 | $0.0245 | −97% | — | — |
| Chutes 0xded6…657c | — | 7.82 | $0.56 | 1.33M | gated | $0.352 | $0.0352 | $2.75 | — | — | — |
| Pearl 0xb4f4…e012 | healthy | 5.41 | $0.02 | 41.57k | gated | $0.20 | $0.05 | $2.00 | — | — | — |
| engySAME 0x9e40…8ec3 | — | 1.88 | $0.32 | 3.18M | gated | $0.045 | $0.015 | $0.32 | −70% | — | — |
| PawFlor-SI-Gateway ⚡SAME 0x70bc…8261 | — | 0.97 | $0.01 | 58.25k | gated | $0.10 | $0.10not advertised | $0.10 | −33% | — | — |
| Fire Ant 🔥🐜SAME 0xbe05…bc5d | — | 0.00 | — | — | gated | $0.423 | $0.0295 | $2.99 | — | — | — |
| LLMartSAME 0x948a…a000 | — | 0.00 | $0.16 | 192.43k | gated | $0.20 | $0.0215 | $2.00 | — | — | — |
| antseed-arctic-otter-aa27 0x4025…aa27 | — | 0.00 | $0.14 | 5.32M | gated | $0.009 | $0.0009 | $0.045 | −94% | — | — |
| ckw 0x2da0…6807 | — | 0.00 | $0.00 | 88.53k | gated | $0.00 | $0.00 | $0.00 | −100% | — | — |
"Best price" and the seller table are live Antseed catalog data (advertised $/1M tokens — or $ per generated image for unit-billed image models — not settled amounts). Reputation = buyer trust score (0-100, the Antseed SDK's own formula). "Routing" = the SDK's default buyer routing (what the VPR desktop app ships with): a trust ≥ 60 gate on the effective reputation, then cheapest-first among routable sellers; live failover state (per-peer cooldowns) is buyer-side runtime and not included. Model knowledge (TLDR, provider, About) via the VeniceStats enrichment layer. Advertised catalog, not the model used in any specific purchase. "Usage on Antseed" counts only settlements whose buyers share the per-model split on-chain (metadata v2/v3, opt-in), so every usage figure is a lower bound.