VerdictPal
Join

Models / Alibaba

Qwen3.8 27B

Qwen · Released 2026-08-14

Alibaba's dense open-weight Qwen3.8 sibling (14 Aug 2026). Artificial Analysis Intelligence Index 52 on the xhigh-effort endpoint. AA hosted list $0.50 / $3 per 1M tokens. Apache 2.0 weights on Hugging Face.

Self-hostable Qwen3.8 dense checkpoint for local and single-GPU runs. Hosted Qwen3.8 Max is a different product. Prefer sourced ledger rows over vendor SWE/OSWorld tables; AA scores sync from the public snapshot.

#22 of 166 on AA Index · current snapshot
52 AA Index · 9 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite60
Context262K
Input / 1M$0.50
Output / 1M$3.0
Knowledge cutoffunpublished
Output speed45 t/s
TTFT (default API)1.21s
Statussolid

Family profile

Best published score in each covered benchmark family.

73/100 avg
Reasoning
91
Coding
45
Agentic
80
Long context
77

4 tested benchmark families

Editor's note
Recorded from the 2026-08-24 AA snapshot (slug qwen3-8-27b, xhigh effort, index 52) plus the Hugging Face card and Alibaba Cloud post. This is the dense 27B checkpoint, not hosted Qwen3.8 Max and not the 2.4T MoE. Vendor SWE-bench Pro / OSWorld figures stay off the ledger until they appear in a mapped source. Editorial depth still needs a desk pass.
Benchmark placements

Where Qwen3.8 27B places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
τ³-Bench BankingAgentic#5/ 10648Artificial Analysis2026-09-01
AA-LCRLong context#25/ 17977Artificial Analysis2026-09-01
Terminal-BenchAgentic#26/ 18480Artificial Analysis2026-09-01
Artificial Analysis Intelligence IndexReasoning#31/ 18452Artificial Analysis2026-09-01
GPQA DiamondReasoning#39/ 18591Artificial Analysis2026-09-01
AA time to first tokenPerformance#57/ 185Artificial Analysis2026-09-01
Humanity's Last ExamReasoning#57/ 18034Artificial Analysis2026-09-01
AA output speedPerformance#73/ 185Artificial Analysis2026-09-01
SciCodeCoding#84/ 18045Artificial Analysis2026-09-01
Family context

The three highest-scoring models with pages in each capability family. Where Qwen3.8 27B shows up, it's highlighted.

ReasoningQwen3.8 27B · 91
  1. 1
    Claude Mythos Preview
    95
  2. 2
    Gemini 3.7 Flash
    95
  3. 3
    Grok 4.6
    95
CodingQwen3.8 27B · 45
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticQwen3.8 27B · 80
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextQwen3.8 27B · 77
  1. 1
    Muse Spark 1.2
    83
  2. 2
    Kimi K3
    83
  3. 3
    Gemini 3.5 Flash
    81
Newest receipts

Benchmark rows added to the public ledger for Qwen3.8 27B in the last 120 days. Older rows live in the full table below.

9 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context77.33%Artificial Analysis2
AA output speedPerformance45 t/sArtificial Analysis2
AA time to first tokenPerformance1.21sArtificial Analysis2
Artificial Analysis Intelligence IndexReasoning52%Artificial Analysis2
GPQA DiamondReasoning90.5%Artificial Analysis2
Humanity's Last ExamReasoning33.9%Artificial Analysis2
SciCodeCoding44.7%Artificial Analysis2
τ³-Bench BankingAgentic48.04%Artificial Analysis2
Terminal-BenchAgentic79.78%Artificial Analysis2
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Reasoning91
Coding45
Agentic80
Long context77
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaMiniMaxxAIZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for Qwen3.8 27B. Scores link to the original source; gaps mean no public row exists yet.

9 sourced rows

9 of 37 catalog benchmarks have a sourced row for Qwen3.8 27B.

Reasoning

Coding

Agentic

Long context

Performance

Changelog
  1. Open weights published. AA Intelligence Index 52 on xhigh; hosted list $0.50 / $3 per 1M tokens.
Sources