Qwen3 30B A3B Thinking 2507

qwen/qwen3-30b-a3b-thinking-2507

Qwen (Alibaba)· 262K ctx· apache-2.0

open weightstool callreasoning quant: unknown

spread (output, p90 ÷ p10)

— too few providers

$1.09 – $1.30 · 2 providers · max÷min 1.2×

2 providers priced · 2 offerings

Including routers & resellers

Cheapest / median / spread use providers that serve the model on their own capacity: first-party lab, hyperscaler, inference host.

Provider classes in the stats:

Blended $/M assumptionsInput : output: 1Cache-hit share of input%

blended = 90.9% input × [(1 − 70% cache) × input + 70% cache × cache-read] + 9.1% × output + cache writes (30% of uncached input) where a write price is published

A provider that publishes no cache price is blended with cache share 0 (all input at the input price) and labelled “no cache price published” — never penalised against identical in/out prices.

Provider	Input	Output	Cache rd	Blended
Qiniu hyperscaler	$0.11	$1.09	—	$0.20no cache price published
Clarifai inference host	$0.36	$1.30	—	$0.45no cache price published

Free tiers (rate-limited) · 1

No payment required; rate limits and quotas apply (not tracked yet). Excluded from cheapest, median and spread.

ModelScope Qwen/Qwen3-30B-A3B-Thinking-2507China-hosted endpoint (modelscope.cn)ModelScope account + access tokenfree tier

Connect

Provider (default: cheapest output among direct providers)

$1.09 /M · hyperscaler

opencode.json

{
  "provider": {
    "qiniu-ai": {
      "npm": "@ai-sdk/openai-compatible",
      "options": {
        "baseURL": "https://api.qnaigc.com/v1"
      },
      "models": {
        "qwen3-30b-a3b-thinking-2507": {}
      }
    }
  }
}

Set QINIU_API_KEY in your environment. Generated — verify against the OpenCode docs.

Output $/M over time

Median line, min–max band. Pre-2026-09 points are catalog-recorded dates from models.dev git history, not provider effective dates.

Data: models.dev (MIT), snapshot 2026-10-22. Prices are vendor-published; verify before purchase. Methodology.