Loading...
Loading...
Qwen3-Coder-Next is an open-weight causal language model optimized for coding agents and local development workflows. It uses a sparse MoE design with 80B total parameters and only 3B activated per token, delivering performance comparable to models with 10 to 20x higher active compute, which makes it well suited for cost-sensitive, always-on agent deployment. The model is trained with a strong agentic focus and performs reliably on long-horizon coding tasks, complex tool usage, and recovery from execution failures. With a native 256k context window, it integrates cleanly into real-world CLI and IDE environments and adapts well to common agent scaffolds used by modern coding tools. The model operates exclusively in non-thinking mode and does not emit <think> blocks, simplifying integration for production coding agents.
Quality Index
28.3
112th of 442
Top 26%
Coding Index
22.9
126th of 352
Top 36%
Price/1M
$0.60
404th cheapest
94% above median
Top 59%
Speed
164 tok/s
Top 10%
TTFT
0.91s
Context Window
262K
61st largest
Top 25%
Input
$0.35
per 1M tokens
Output
$1.20
per 1M tokens
Blended
$0.60
per 1M tokens
Cheaper than 41% of models. Median price is $0.31/1M tokens.
Daily
$0.60
Monthly
$18.00
164
tokens/sec
Faster than 90% of models
0.91
seconds
Faster than 33% of models
0.91
seconds
Faster than 40% of models
Market Median
46 tok/s
260% faster
Median TTFT
0.42s
117% slower
Throughput/Dollar
273
tok/s per $/1M
Speed Comparison
Context Window
262K
tokens
Larger than 75% of models
Max Output
66K
tokens
25% of context