modelbenchmark.io

Qwen3 235B-A22B Thinking 2507 (Deep Infra)

Alibaba · qwen3-235b-a22b-thinking

Compare

Updated large open Qwen3 MoE thinking model for reasoning, math, coding, and tool use

Specification

most-agreed values

Context
262K
Max output
82K
Released
2025-07-25
Knowledge cutoff
2024-12
Retires
—
Open weights
yes
Input
text
Output
text

Price

US dollars per million tokens · most-agreed

Input
$0.23
Output
$2.30
Cache read
$0.20
Cache write
$0.22

Available from 15 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
modelscopefreefree——262K131K—
io-net$0.11$0.60$0.055$0.22262K4K—
submodel$0.20$0.60——262K131K—
fireworks_ai$0.22$0.88——262K262K—
edenai$0.23$2.30$0.20—262K82K—
qiniu-ai$0.2899$2.8986——262K4K—
deepinfra$0.30$2.90——262K262K—
helicone$0.30$2.90——262K82K—
huggingface$0.30$3.00——262K82K—
jiekou$0.30$3.00——131K131K—
nano-gpt$0.30$0.50——131K82K—
novita$0.30$3.00——131K33K—
kilo$0.45$3.50——128K16K—
openrouter$0.45$3.50——128K16K—
venice$0.45$3.50——128K16K—

Price history

append-only observations · a listing writes a row only when its price moves

DateHostField$/M
2026-09-02kiloInput$0.23
2026-10-10kiloInput$0.45
2026-09-02nano-gptInput$0.30
2026-10-22nano-gptInput$0.30
2026-09-02openrouterInput$0.23
2026-09-06openrouterInput$0.23
2026-10-10openrouterInput$0.45
2026-09-02kiloOutput$2.30
2026-10-10kiloOutput$3.50
2026-09-02nano-gptOutput$0.50
2026-10-22nano-gptOutput$0.50
2026-09-02openrouterOutput$2.30
2026-09-06openrouterOutput$2.30
2026-10-10openrouterOutput$3.50

More from Alibaba

most-hosted first