modelbenchmark.io

Qwen3 8B

Alibaba · qwen3-8b

Compare

Qwen instruction model for multilingual chat, reasoning, and tool use

Specification

alibaba listing

Context
131K
Max output
8K
Released
2025-04
Knowledge cutoff
2025-04
Retires
—
Open weights
yes
Input
text
Output
text

Price

US dollars per million tokens · alibaba list price

Input
$0.18
Output
$0.70
Cache read
$0.235

Quality

2 benchmarks · 1 source

Composite
48th
percentile of 344 scored models
Rank
181 / 344
±0.09 sd
Evidence
2 × 1
too little evidence to place with confidence
Effort range
—
one configuration only
reasoning43rd1/4 bench
math57th1/7 bench

Every score

one row per source and configuration — nothing averaged away

BenchmarkScoreConfigurationSourceRun
OTIS Mock AIME 2024-202556.1 ±6.5defaultEpoch AI2026-08-27
GPQA diamond56.8 ±2.9defaultEpoch AI2026-08-27

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

Available from 9 hosts

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
alibabavendor$0.18$0.70——131K8K—
siliconflow-cnfreefree——131K131K—
llamagate · llamagate$0.04$0.14——33K8K—
siliconflow$0.06$0.06——131K131K—
huggingface$0.07$0.18——41K33K—
alibaba-cn$0.072$0.287——131K8K—
thinkingmachines$0.195$0.60$0.039—33K33K—
fireworks_ai$0.20$0.20——41K41K—
nano-gpt$0.47$0.47$0.235—41K33K—

Price history

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

DateHostField$/M
2026-09-02openrouterInput$0.117
2026-09-06openrouterInput$0.117
2026-09-02siliconflow-cnInput$0.06
2026-10-22siliconflow-cnInputfree
2026-09-02openrouterOutput$0.455
2026-09-06openrouterOutput$0.455
2026-09-02siliconflow-cnOutput$0.06
2026-10-22siliconflow-cnOutputfree
2026-09-05nano-gpt · teeCache read$0.11
2026-09-09nano-gpt · teeCache read$0.11
2026-09-05nano-gpt · teeInput$0.11
2026-09-09nano-gpt · teeInput$0.11
2026-09-05nano-gpt · teeOutput$0.45
2026-09-09nano-gpt · teeOutput$0.45

More from Alibaba

most-hosted first