llm.ing

Qwen3.8 27B

Liquid Inferenceopen weightsOpenRouter ↗Hugging Face ↗Artificial Analysis ↗

Dense 27B vision-language model for coding, agent tasks, and image and video understanding

Benchmarks

BenchmarkCategoryScoreVariantProvenanceSourceObserved
AA Agentic Indexagentic45.8—mirroropenrouterNov 15, 2026
AA Coding Indexcoding68.1—mirroropenrouterNov 15, 2026
AA Intelligence Indexintelligence20.1—mirroropenrouterNov 15, 2026
10.0non-reasoningindependentaaNov 15, 2026
LMArena Visionpreference1270—crowdlmarenaNov 9, 2026
LMArena WebDevpreference1590—crowdlmarenaNov 11, 2026
Median output speedspeed50 tok/snon-reasoningindependentaaNov 19, 2026
Median time to first tokenspeed3.85snon-reasoningindependentaaNov 19, 2026

Includes index data from Artificial Analysis.

Providers

Live operational stats per hosting endpoint, via OpenRouter.

ProviderQuantContext$/1M in · outUptime 24hLatencyThroughput
AkashMLfp8262K$0.225 · $1.98100.00%——
Alibaba—1M$0.425 · $2.5599.83%——
Cerebrasfp1666K$0.99 · $1.49100.00%——
Chutesfp8262K$0.039 · $1.2599.93%——
Cloudflare—262K$0.45 · $3.294.44%——
CoreWeavefp8262K$0.4 · $399.85%——
Darkbloomfp4262K$0.035 · $1.598.37%——
DeepInfrafp8262K$0.15 · $1.88100.00%——
DekaLLM—262K$0.04 · $2.5100.00%——
GMICloudfp81M$0.45 · $3.285.71%——
Ionstreamfp8262K$0.04 · $1.9599.59%——
Mancer 2fp8262K$0.125 · $1.585.98%——
Modal—262K$0.04 · $2.3100.00%——
ModelRunfp4262K$0.2 · $2.5100.00%——
Near AIfp8262K$0.04 · $1.35100.00%——
Nebiusfp41M$0.45 · $3100.00%——
Novita—1M$0.42 · $3100.00%——
Parasailfp8262K$0.24 · $2.2100.00%——
PrimeIntellectfp8262K$0.402 · $2.8699.82%——
Rekafp8262K$0.15 · $2.5100.00%——
SiliconFlowfp8262K$0.425 · $2.55100.00%——
StreamLake—1.0M$0.175 · $1.05100.00%——
Tenstorrentfp8262K$0.15 · $1.699.48%——
Venicefp8262K$0.45 · $3.2100.00%——
Waferfp8262K$0.06 · $2.299.65%——