Powered byLightning AI
← Managed inference

Qwen3 235B Instruct

Highest-quality answers when accuracy matters more than cost

Start free

Overview

Qwen3-235B-A22B-Instruct-2507 is Qwen's largest open-weight instruct model: 235B total parameters with 22B active per token. The 2507 refresh targeted instruction following, logical reasoning, maths, science, coding and tool use, added long-tail knowledge across multiple languages, and improved 256K long-context understanding. It runs in non-thinking mode only.

Strengths

  • Broadest general knowledge of the chat models
  • 262K native context
  • Strong multilingual coverage

Use cases

  • Hard reasoning or analysis where quality dominates
  • Agents that need the best tool-use accuracy
  • Synthesis across many long documents

Good to know

Non-thinking mode only, so it emits no reasoning blocks. At 22B active parameters per token, expect the highest latency and cost of the chat models here.

API access

One OpenAI-compatible API for every model in the catalog, plus the CLI and the function SDK inside workflows. Sign up free, add a card for $5 in credits, and these requests work:

qwen3-235b · curl
curl https://gateway.graphn.ai/v1/chat/completions \
  -H "Authorization: Bearer $GRAPHN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen3-235b",
    "messages": [{"role": "user", "content": "Hello"}]
  }'