Qwen3-8B

reasoning

Qwen3-8B represents a new tier in the series. With 8.2 billion parameters, this model retains an architecture of 36 layers and 32 attention heads, but introduces several important improvements — it no longer uses tied embeddings, and its context window has been expanded to 40K tokens, enabling excellent capabilities for handling long documents and complex tasks.

The doubling of parameter count compared to the 4B version significantly enhances response quality across all task types, especially in mathematical reasoning, programming, and advanced analysis. The model excels in tasks requiring multi-step reasoning and deep contextual understanding. Built-in support for both *thinking* and *non-thinking* modes allows performance optimization based on task complexity and available processing time, while the *Thinking Budget* mechanism enables fine-grained control over computational intensity for optimal efficiency.

Qwen3-8B is ideal for advanced professional applications such as financial analysis, medical diagnostics, and legal practice. It is well-suited for building intelligent assistants for professionals, automated technical documentation systems, and educational platforms.


Announce Date: 29.04.2025
Parameters: 9B
Context: 41K
Layers: 36
Attention Type: Full or Sliding Window Attention
Developer: Qwen
Transformers Version: 4.51.0
License: Apache 2.0

Public endpoint

Use our pre-built public endpoints for free to test inference and explore Qwen3-8B capabilities. You can obtain an API access token on the token management page after registration and verification.
Model Name Context Type GPU Status Link
There are no public endpoints for this model yet.

Private server

Rent your own physically dedicated instance with hourly or long-term monthly billing.

We recommend deploying private instances in the following scenarios:

  • maximize endpoint performance,
  • enable full context for long sequences,
  • ensure top-tier security for data processing in an isolated, dedicated environment,
  • use custom weights, such as fine-tuned models or LoRA adapters.

Recommended server configurations for hosting Qwen3-8B

Prices:
Name GPU Price, hour TPS Max Concurrency
teslaa10-1.16.32.160
40,960.0
 1 $0.53 2.084 Launch
teslat4-2.16.32.160
40,960.0
tensor
 2 $0.54 2.580 Launch
teslaa2-2.16.32.160
40,960.0
tensor
 2 $0.57 2.593 Launch
rtx2080ti-2.12.64.160
40,960.0
tensor
 2 $0.69 1.271 Launch
rtx3090-1.16.24.160
40,960.0
 1 $0.83 2.257 Launch
rtx4090-1.16.32.160
40,960.0
 1 $1.02 2.250 Launch
rtxa5000-2.16.64.160.nvlink
40,960.0
tensor
 2 $1.23 5.177 Launch
rtx3080-3.16.64.160
40,960.0
pipeline
 3 $1.43 1.946 Launch
rtx5090-1.16.64.160
40,960.0
 1 $1.59 3.531 Launch
rtx3080-4.16.64.160
40,960.0
tensor
 4 $1.82 2.931 Launch
teslaa100-1.16.64.160
40,960.0
 1 $2.37 11.365 Launch
h100-1.16.64.160
40,960.0
 1 $3.83 11.354 Launch
h100nvl-1.16.96.160
40,960.0
 1 $4.11 13.629 Launch
teslaa100-2.24.96.160.nvlink
40,960.0
tensor
 2 $4.61 23.740 Launch
h200-1.16.128.160
40,960.0
 1 $4.74 21.269 Launch
h200-2.24.256.160.nvlink
40,960.0
tensor
 2 $9.40 43.547 Launch
Prices:
Name GPU Price, hour TPS Max Concurrency
teslaa10-1.16.32.160
40,960.0
 1 $0.53 1.531 Launch
teslat4-2.16.32.160
40,960.0
tensor
 2 $0.54 2.027 Launch
teslaa2-2.16.32.160
40,960.0
tensor
 2 $0.57 2.040 Launch
rtx3090-1.16.24.160
40,960.0
 1 $0.83 1.704 Launch
rtx2080ti-3.12.24.120
40,960.0
pipeline
 3 $0.84 1.859 Launch
rtx4090-1.16.32.160
40,960.0
 1 $1.02 1.698 Launch
rtx2080ti-4.16.32.160
40,960.0
tensor
 4 $1.12 3.000 Launch
rtxa5000-2.16.64.160.nvlink
40,960.0
tensor
 2 $1.23 4.624 Launch
rtx3080-3.16.64.160
40,960.0
pipeline
 3 $1.43 1.393 Launch
rtx5090-1.16.64.160
40,960.0
 1 $1.59 2.978 Launch
rtx3080-4.16.64.160
40,960.0
tensor
 4 $1.82 2.378 Launch
teslaa100-1.16.64.160
40,960.0
 1 $2.37 10.813 Launch
h100-1.16.64.160
40,960.0
 1 $3.83 10.801 Launch
h100nvl-1.16.96.160
40,960.0
 1 $4.11 13.076 Launch
teslaa100-2.24.96.160.nvlink
40,960.0
tensor
 2 $4.61 23.188 Launch
h200-1.16.128.160
40,960.0
 1 $4.74 20.716 Launch
h200-2.24.256.160.nvlink
40,960.0
tensor
 2 $9.40 42.994 Launch
Prices:
Name GPU Price, hour TPS Max Concurrency
teslat4-3.32.64.160
40,960.0
pipeline
 3 $0.88 2.672 Launch
teslaa10-2.16.64.160
40,960.0
tensor
 2 $0.93 3.474 Launch
teslat4-4.16.64.160
40,960.0
tensor
 4 $0.96 4.467 Launch
teslaa2-3.32.128.160
40,960.0
pipeline
 3 $1.06 2.691 Launch
rtx2080ti-4.16.32.160
40,960.0
tensor
 4 $1.12 1.850 Launch
rtxa5000-2.16.64.160.nvlink
40,960.0
tensor
 2 $1.23 3.474 Launch
teslaa2-4.32.128.160
40,960.0
tensor
 4 $1.26 4.493 Launch
rtx3090-2.16.64.160
40,960.0
tensor
 2 $1.56 3.821 Launch
rtx5090-1.16.64.160
40,960.0
 1 $1.59 1.829 Launch
rtx3080-4.16.64.160
40,960.0
tensor
 4 $1.82 1.228 Launch
rtx4090-2.16.64.160
40,960.0
tensor
 2 $1.92 3.808 Launch
teslaa100-1.16.64.160
40,960.0
 1 $2.37 9.663 Launch
h100-1.16.64.160
40,960.0
 1 $3.83 9.651 Launch
h100nvl-1.16.96.160
40,960.0
 1 $4.11 11.926 Launch
teslaa100-2.24.96.160.nvlink
40,960.0
tensor
 2 $4.61 22.038 Launch
h200-1.16.128.160
40,960.0
 1 $4.74 19.566 Launch
h200-2.24.256.160.nvlink
40,960.0
tensor
 2 $9.40 41.845 Launch

Related models

Qwen3-30B-A3B
Qwen3-235B-A22B
Qwen3-0.6B
Qwen3-1.7B
Qwen3-4B
Qwen3-14B
Qwen3-32B
DeepSeek-R1-0528-Qwen3-8B
T-pro-it-2.0
Qwen3-235B-A22B-Instruct-2507
Qwen3-235B-A22B-Thinking-2507
Qwen3-Coder-480B-A35B-Instruct
Qwen3-Coder-30B-A3B-Instruct
Qwen3-30B-A3B-Instruct-2507
Qwen3-30B-A3B-Thinking-2507
Qwen3-4B-Instruct-2507
Qwen3-4B-Thinking-2507
Qwen3-VL-235B-A22B-Instruct
Qwen3-VL-235B-A22B-Thinking
Qwen3-VL-30B-A3B-Instruct
Qwen3-VL-30B-A3B-Thinking
Qwen3-VL-8B-Instruct
Qwen3-VL-8B-Thinking
Qwen3-VL-4B-Instruct
Qwen3-VL-4B-Thinking
Qwen3-VL-2B-Instruct
Qwen3-VL-2B-Thinking
Qwen3-VL-32B-Instruct
Qwen3-VL-32B-Thinking
avibe

Need help?

Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.