Qwen3-Next 80B-A3B (Thinking)
Efficient Qwen thinking model for local reasoning, math, and coding agents
Context
262,144
Max output
262,000
Input / 1M
$0.14
Output / 1M
$0.65
Providers
16
01Profile
Profile
Released
2025-09
Last updated
2026-01
Knowledge cutoff
2025-04
Weights
Open
02Capabilities
Capabilities
Input
text
Output
text
Reasoning
Yes
Tool calling
Yes
Structured output
No
Attachments
No
Cache read
Not listed
Cache write
Not listed
03Pricing
Pricing across 15 providers
Per 1M tokens, USD. Sorted by input price. First-party row highlighted.
04History
Change history
12+ recent events since 2025-09
2026-09-15ContextkiloContext window262,144→ new
2026-09-15ContextkiloMax output235,929→ new
2026-09-15ContextopenrouterMax output235,929→ new
2026-09-11Price upamazon-bedrockinput / 1M$0.14→ $0.15
2026-09-11Price cutamazon-bedrockoutput / 1M$1.40→ $1.20
2026-09-11Contextamazon-bedrockContext window262,000→ 262,144
2026-09-07Removedcortecsofferingremoved→ new
2026-09-07ContextkiloContext window—→ 262,144
2026-09-07ContextkiloMax output—→ 235,929
2026-09-07ContextopenrouterMax output—→ 235,929
2026-09-02Removednebiusofferingremoved→ new
2026-08-31Removedllmgateway-providersofferingremoved→ new