Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) vs Qwen3 Coder Next
Which is better for the money: Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) by Anthropic or Qwen3 Coder Next by Alibaba? Intelligence, total benchmark cost, and end-to-end time from Artificial Analysis & other sources, snapshot 2026-09-09.
Depends on your budget
Claude Sonnet 4.6 (Adaptive Reasoning, … scores 20.4 points higher (30.5 vs 10.1) but costs 10× as much to run the benchmark ($4925 vs $476; flagship vs mid-range cost classes). They are not alternatives at one budget: pick Claude Sonnet 4.6 (Adaptive Reasoning, … when you need intelligence above 10.1, and Qwen3 Coder Next when 10.1 is enough.
| Metric | Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | Qwen3 Coder Next | Difference |
|---|---|---|---|
| Intelligence (higher is better) | 30.5 | 10.1 | −66.9% Qwen3 Coder Next is 66.9% less intelligent |
| Total AA cost (lower is better) | $4,924.80 | $476.03 | −90.3% Qwen3 Coder Next is 90.3% cheaper |
| End-to-end time (lower is better) | 141.12s | 5.52s | −96.1% Qwen3 Coder Next is 96.1% faster |
| Price per 1M tokens (in / out) | $3.00 / $15.0 | $0.35 / $1.20 | |
| Input modalities | Not reported | Not reported |
Qwen3 Coder Next vs Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort): 66.9% less intelligence, 90.3% less cost to run the AA benchmark, and 96.1% faster end-to-end.
Against the frontier
- Claude Sonnet 4.6 (max) is off the frontier: at the same or lower cost, GPT-6 Astra (xhigh) scores 52.50 for $3803.
- Qwen3 Coder Next is off the frontier: at the same or lower cost, GLM-5.3-Flash scores 41.90 for $280.
Compare something else
Accept: text/markdown or ?format=md returns this page as Markdown. JSON: /api/v1/compare/claude-sonnet-4-6-adaptive/qwen3-coder-next · Markdown: /api/v1/compare/claude-sonnet-4-6-adaptive/qwen3-coder-next?format=md · llms.txt · agent skill