Skip to content
← Leaderboard

Qwen3.8 27B

Open source
Self-hostable
Free to run
Alibaba (Qwen)
Open license
text
vision
Released 2mo ago
Avg score
57.4
/ 100 · not enough data for the index
Context
1.0M
Output limit
236k
Input price
$0.43 /M
Output price
$2.55 /M
Parameters
27.8B
Est. VRAM (Q4)
~16.7 GB

Price from the model maker's own API, verified 8h ago · cached input $0.09 /M

Added automatically from the public model catalog. Family and licence details are not curated yet; benchmark scores appear as upstream leaderboards list this model.

Benchmarks

preference

Chatbot Arena EloFresh
Elo

Crowdsourced pairwise human preference rankings of LLM responses. Higher Elo means more frequently preferred by users.

performance

Output SpeedN/A
tok/s

Median output speed in tokens per second over the last 14 days on the model maker's own API, measured nightly with a fixed ~800-token prompt and ~250-token answer. Models not yet probed fall back to a dated manual figure. Higher is faster.

Time to First TokenN/A
ms

Median time from request to first streamed token in milliseconds over the last 14 days on the model maker's own API, using the same nightly probe prompt. Lower is snappier; reasoning models are penalised here because they think before talking.

reasoning

ARC-AGI 2Fresh
%

Second-generation ARC challenge testing fluid reasoning over abstract visual puzzles. Resists training-data memorisation by construction: each puzzle is novel and solutions require multi-step pattern induction. Frontier models are only just starting to score above chance on the harder tier.

reliability

API Success RateN/A
%

Share of nightly probe requests over the last 14 days that streamed a complete answer without an error, timeout or empty response. Needs at least five attempts before a figure is shown.

Reliability monitor

Loading drift signal…

Hosted endpoints

HostInput $/MOutput $/MContextSpeedTTFTUptime 24hQuant
Host 148$0.02$0.50262k35 tok/s0.60 s89.6%fp4
Host 149$0.04$1.25262k49 tok/s1.68 s99.9%fp8
Host 124$0.04$1.35262k——99.7%fp8
Host 155$0.99$1.4966k——100.0%fp16
Host 153$0.15$1.88262k——98.9%bf16
Host 150$0.04$1.95262k——99.6%fp8
Host 28$0.23$1.98262k——97.6%fp8
Host 106$0.04$2.20262k——99.5%fp4
Host 151$0.04$2.20262k——100.0%fp8
Host 29$0.24$2.20262k——99.7%fp8
Host 82$0.04$2.30262k——100.0%unknown
Host 152$0.06$2.30262k——100.0%fp8
Host 138$0.04$2.50262k——100.0%unknown
Host 116$0.18$2.50262k——99.4%fp8
Host 62$0.20$2.50262k——99.5%fp4
Host 86$0.20$2.501.0M——99.5%unknown
Host 90maker's API$0.43$2.551.0M62 tok/s1.02 s99.4%unknown
Host 115$0.40$3.00262k——99.8%fp8
Host 154$0.42$3.001.0M——98.7%unknown
Host 71$0.45$3.20262k——100.0%unknown
Host 52$0.45$3.20262k——100.0%fp8
Anonymised third-party hosts. Sorted by lowest output price. Speed and time to first token are medians from our nightly probe where measured; uptime is the catalog's 24-hour figure.

Compare with...