All AI models · LLM benchmarks · Methodology

Alibaba

Qwen3.8 2.4T A95B

Open-weight sparse MoE (2.4T total, 95B active), the open-weight twin of Qwen3.8 Max for coding, research, complex reasoning, and agentic workflows

Model facts

Context window
262K tokens
Maximum output
262K tokens
Cheapest paid input
$2 per 1M tokens
Cheapest paid output
$6 per 1M tokens
Open weights
yes
Providers
15

Capabilities and modalities

reasoning, tool calling, structured output, attachments, temperature control, open weights, text, image, video, pdf.

Local hardware estimate

At 4K context: Q4_K_M 1412.6 GiB, Q8_0 2551.7 GiB, F16 4687.5 GiB working memory.

Parameters
2446.2 billion
Architecture
qwen3_5moe
Model license
other

Planning estimate adapted from the MIT-licensed whichllm estimator using metadata from Hugging Face; not a vendor minimum requirement.

Artificial Analysis benchmarks

Compare this model on the full LLM benchmark leaderboard.

ConfigurationAAICodingMathOutput tokens/s
default57.771.9not measured40.2

Observed price history

  • 2026-08-24: $1.8 input / $5.4 output per 1M tokens
  • 2026-08-28: $2 input / $6 output per 1M tokens

API providers