modelbenchmark.io

Qwen3.5 397B-A17B

Alibaba · qwen3-5-397b-a17b

Compare

Large open Qwen multimodal MoE for visual agents and long technical tasks

Specification

alibaba listing

Context
262K
Max output
66K
Released
2026-02-15
Knowledge cutoff
2025-04
Retires
2026-06-29
Open weights
yes
Input
text, image, video, audio
Output
text

Price

US dollars per million tokens · alibaba list price

Input
$0.60
Output
$3.60
Cache read
$0.22
Cache write
$0.75

Quality

4 benchmarks · 1 source

Composite
70th
percentile of 334 scored models
Rank
100 / 334
±0.08 sd
Evidence
4 × 1
one source, or fewer than 3 benchmarks
Effort range
one configuration only
reasoning75th2/4 bench
math68th2/7 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
OTIS Mock AIME 2024-202588.9 ±4.7defaultEpoch AI2026-08-07
GPQA diamond85.9 ±2.5defaultEpoch AI2026-08-07
Chess Puzzles12.0 ±3.3defaultEpoch AI2026-08-07
FrontierMath-Tiers-1-3-v229.5 ±2.7defaultEpoch AI2026-08-28

The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.

Available from 38 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
alibabavendor$0.60$3.60262K66K
nvidiafreefree262K8K
unorouter · freefreefree262K66K
alibaba-cn$0.172$1.032262K66K
empiriolabs$0.172$1.032$0.172256K64K
merge-gateway$0.172$1.032$0.0344256K64K
orcarouter$0.172$1.032262K66K
siliconflow-cn$0.29$1.74262K66K
vultr$0.30$2.00262K66K
crof$0.35$1.75$0.07262K262K
kilo$0.39$2.34262K236K
siliconflow$0.39$2.34262K262K
tokengo$0.40$2.65$0.20262K66K
deepinfra$0.45$3.00$0.22262K82K
digitalocean$0.55$3.50$0.111131K131K
nano-gpt · tee$0.55$3.50$0.275262K66K
ofox$0.55$3.50$0.55256K64K
ofox · bailian$0.55$3.50$0.55256K64K
openrouter$0.55$3.50$0.225262K
236K
66K
cloudflare-ai-gateway$0.60$3.60262K66K
huggingface$0.60$3.60262K33K
jalapeno$0.60$3.60262K66K
llmtr$0.60$3.60256K66K
mixlayer$0.60$3.60262K262K
nano-gpt$0.60$3.60$0.30258K66K
nano-gpt · thinking$0.60$3.60$0.30258K66K
nebius$0.60$3.60$0.06$0.75262K
8K
262K
novita$0.60$3.60262K66K
novita-ai$0.60$3.60262K64K
scaleway$0.60$3.60256K16K
scaleway · scaleway$0.60$3.60256K16K
together_ai$0.60$3.60$0.35262K2026-06-29
togetherai$0.60$3.60$0.35262K130K
cortecs$0.668$4.01262K262K
ovhcloud$0.71$4.25262K262K
venice$0.75$4.50128K33K
greenpt$0.798$4.959262K16K
qiniu-ai256K64K

Price history

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

DateHostField$/M
2026-09-02nebiusInput$0.60
2026-09-06nebiusInput$0.60
2026-09-02openrouterInput$0.39
2026-09-03openrouterInput$0.55
2026-09-06openrouterInput$0.55
2026-09-07openrouterInput$0.39
2026-09-09openrouterInput$0.55
2026-09-02nebiusOutput$3.60
2026-09-06nebiusOutput$3.60
2026-09-02openrouterOutput$2.34
2026-09-03openrouterOutput$3.50
2026-09-06openrouterOutput$3.50
2026-09-07openrouterOutput$2.34
2026-09-09openrouterOutput$3.50
2026-09-03openrouterCache read$0.225
2026-09-09openrouterCache read$0.225

More from Alibaba

most-hosted first