Alibaba Qwen: Qwen3.8 27B

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

Source-linked qwen Open weights Released 2026-08-14

API record Report

InputT

OutputT

Input price$0.04/M

Output price$1.35/M

Context262.144K

Max output32.768K

Providers51

Inference availability

Providers

Provider-specific identifiers, limits, and listed prices per million tokens. Every row links back to the provider's own documentation.

Report a price

Providers offering Qwen3.8 27B
Provider	Provider model ID	Context	Max output	Input	Output	Cache read	Capabilities	Docs
AMD	`Qwen3.8-27B`	262.144K	32.768K	—	—	—	ReasoningToolsJSON	Docs ↗
IteraCompute	`qwen/qwen3.8-27b`	131.072K	8.192K	—	—	—	ReasoningToolsJSON	Docs ↗
Blue Claw	`Qwen3.8-27B`	131.072K	32.768K	—	—	—	ReasoningToolsJSON	Docs ↗
NEAR AI Cloud	`Qwen/Qwen3.8-27B`	262.144K	32.768K	$0.04	$1.35	$0.018	ReasoningToolsJSON	Docs ↗
engy	`qwen3.8-27b`	1.00154M	65.536K	$0.045	$0.32	$0.015	ReasoningToolsJSON	Docs ↗
DevPass (LLM Gateway)	`Qwen3.8-27B`	262K	32.768K	$0.08	$0.35	$0.05	ReasoningTools	Docs ↗
RunInfra	`Qwen/Qwen3.8-27B`	262.144K	32.768K	$0.1	$0.4	$0.01	ReasoningToolsJSON	Docs ↗
Cortecs	`qwen3.8-27b`	1M	500K	$0.1	$0.4	$0.04	ReasoningToolsJSON	Docs ↗
NanoGPT	`qwen3.8-27b`	262.144K	32.768K	$0.15	$0.7	$0.04	ReasoningToolsJSON	Docs ↗
NanoGPT	`qwen/qwen3.8-27b`	1.04858M	32.768K	$0.15	$0.7	$0.04	ReasoningToolsJSON	Docs ↗
Vultr	`qwen3.8-27b`	262.144K	32.768K	$0.15	$1	—	ReasoningToolsJSON	Docs ↗
EmpirioLabs AI	`qwen3-8-27b`	262.144K	32.768K	$0.17	$0.5	$0.08	ReasoningToolsJSON	Docs ↗
OpenRouter	`qwen/qwen3.8-27b`	1M	235.929K	$0.2	$2.5	$0.09	ReasoningToolsJSON	Docs ↗
Deep Infra	`Qwen/Qwen3.8-27B`	262.144K	32.768K	$0.2	$2.5	$0.05	ReasoningToolsJSON	Docs ↗
CrofAI	`qwen3.8-27b`	262.144K	262.144K	$0.2	$1.5	$0.03	ReasoningToolsJSON	Docs ↗
LLM Tech	`unsloth/Qwen3.8-27B-NVFP4`	262.144K	32.768K	$0.25	$2.09	$0.04	ReasoningToolsJSON	Docs ↗
LLM Tech	`nvidia/Qwen3.8-27B-NVFP4`	262.144K	32.768K	$0.25	$2.09	$0.04	ReasoningToolsJSON	Docs ↗
IteraCompute	`iteracompute/qwen3.8-27b`	327.68K	65.536K	$0.3	$2.5	—	ReasoningToolsJSON	Docs ↗
Ambient	`qwen/qwen3.8-27b`	32.768K	8.192K	$0.32	$3.2	$0.16	ReasoningToolsJSON	Docs ↗
OrcaRouter	`qwen/qwen3.8-27b`	262.144K	32.768K	$0.33	$2.4	—	ReasoningToolsJSON	Docs ↗
AKI.IO	`qwen3.8-27b`	262.144K	32.768K	$0.336	$2.467	$0.112	ReasoningToolsJSON	Docs ↗
Kosmik Compute	`qwen/qwen3.8-27b`	262.144K	32.768K	$0.35	$2.2	$0.09	ReasoningToolsJSON	Docs ↗
Hugging Face	`Qwen/Qwen3.8-27B`	262.144K	32.768K	$0.4	$3	—	ReasoningToolsJSON	Docs ↗
ai&	`qwen/qwen3.8-27b`	262.144K	32.768K	$0.4	$3	$0.2	ReasoningToolsJSON	Docs ↗
CoreWeave	`Qwen/Qwen3.8-27B`	262.144K	262.144K	$0.4	$3	$0.15	ReasoningToolsJSON	Docs ↗
TensorX	`qwen/qwen3.8-27b`	262.144K	32.768K	$0.4	$2.4	$0.1	ReasoningToolsJSON	Docs ↗
CommandCode	`Qwen/Qwen3.8-27B`	262.144K	32.768K	$0.4	$3	$0.04	ReasoningToolsJSON	Docs ↗
DevPass (LLM Gateway)	`qwen3.8-27b`	1M	32.768K	$0.42	$3	$0.085	ReasoningToolsJSON	Docs ↗
Novita AI	`qwen/qwen3.8-27b`	1M	131.072K	$0.42	$3	$0.085	ReasoningToolsJSON	Docs ↗
Kilo Gateway	`qwen/qwen3.8-27b`	262.144K	235.929K	$0.425	$2.55	$0.085	ReasoningToolsJSON	Docs ↗
Standard Compute	`qwen/qwen3.8-27b`	262.144K	32.768K	$0.425	$2.55	$0.085	ReasoningToolsJSON	Docs ↗
Neuralwatt	`qwen-3.8-27b`	262.128K	131.072K	$0.45	$3.2	$0.25	ReasoningToolsJSON	Docs ↗
Cloudflare Workers AI	`@cf/qwen/qwen3.8-27b`	262.144K	262.144K	$0.45	$3.2	$0.05	ReasoningToolsJSON	Docs ↗
Ofox	`bailian/qwen3.8-27b`	1.13107M	131.072K	$0.45	$3.2	$0.05	ReasoningToolsJSON	Docs ↗
Nebius Token Factory	`Qwen/Qwen3.8-27B`	262.144K	262.144K	$0.45	$3	$0.1	ReasoningToolsJSON	Docs ↗
Berget.AI	`Qwen/Qwen3.8-27B-FP8`	262.144K	65.536K	$0.46	$3.48	—	ReasoningToolsJSON	Docs ↗
AI/ML API	`alibaba/qwen3.8-27b`	262.144K	131.072K	$0.476	$3.386	—	ReasoningToolsJSON	Docs ↗
Vercel AI Gateway	`alibaba/qwen3.8-27b`	1M	131.072K	$0.5	$3	$0.1	ReasoningToolsJSON	Docs ↗
Charm Hyper	`qwen3.8-27b`	1M	128K	$0.5	$3	$0.1	ReasoningToolsJSON	Docs ↗
Eden AI	`qwen/qwen3.8-27b`	1M	32.768K	$0.5	$3	$0.1	ReasoningToolsJSON	Docs ↗
Ofox	`qwen/qwen3.8-27b`	1.13107M	131.072K	$0.5	$1.71	$0.043	ReasoningToolsJSON	Docs ↗
Regolo AI	`qwen3.8-27b`	120K	120K	$0.58	$2.42	—	ReasoningToolsJSON	Docs ↗
Opper	`qwen3.8-27b`	262.144K	32.768K	$0.581	$3	—	ReasoningTools	Docs ↗
Scaleway	`qwen3.8-27b`	262.144K	32.768K	$0.684	$3.762	$0.137	ReasoningToolsJSON	Docs ↗
Thinking Machines	`Qwen/Qwen3.8-27B:peft:262144`	262.144K	262.144K	$0.698	$5.595	$0.14	ReasoningTools	Docs ↗
Thinking Machines	`Qwen/Qwen3.8-27B`	65.536K	65.536K	$0.698	$5.595	$0.14	ReasoningTools	Docs ↗
Groq	`qwen/qwen3.8-27b`	131.072K	16.384K	$0.8	$4	—	ReasoningToolsJSON	Docs ↗
Tempr Gateway	`groq/qwen/qwen3.8-27b`	131.042K	16.384K	$0.8	$4	$0.4	ReasoningToolsJSON	Docs ↗
Cerebras	`qwen-3.8-27b`	131.072K	40.96K	$0.99	$1.49	$0.99	ReasoningToolsJSON	Docs ↗
Tempr Gateway	`cerebras/qwen-3.8-27b`	131.072K	40.96K	$0.99	$1.49	—	ReasoningToolsJSON	Docs ↗
Prime Inference	`qwen/qwen3.8-27b`	1M	235.929K	$0.99	$3.2	—	ReasoningToolsJSON	Docs ↗

Capability badges appear only where the provider catalog explicitly lists support. A blank cell means the source is silent, not that the feature is absent.

Listed rates

Price across providers

Input price per million tokens as published by each provider. Bars are drawn from listed rates only — no traffic weighting, since the catalog observes no requests.

Lowest input $0.04/M

Across 48 priced providers

Median input $0.41/M

Midpoint of listed rates

Highest input $0.99/M

24.8× the lowest listed rate

Output range $0.32 – $5.595

Per million output tokens

NEAR AI Cloud $0.04/MLowest

engy $0.045/M

DevPass (LLM Gateway) $0.08/M

RunInfra $0.1/M

Cortecs $0.1/M

NanoGPT $0.15/M

Vultr $0.15/M

EmpirioLabs AI $0.17/M

OpenRouter $0.2/M

Deep Infra $0.2/M

CrofAI $0.2/M

LLM Tech $0.25/M

IteraCompute $0.3/M

Ambient $0.32/M

OrcaRouter $0.33/M

AKI.IO $0.336/M

Kosmik Compute $0.35/M

Hugging Face $0.4/M

ai& $0.4/M

CoreWeave $0.4/M

TensorX $0.4/M

CommandCode $0.4/M

DevPass (LLM Gateway) $0.42/M

Novita AI $0.42/M

Kilo Gateway $0.425/M

Standard Compute $0.425/M

Neuralwatt $0.45/M

Cloudflare Workers AI $0.45/M

Ofox $0.45/M

Nebius Token Factory $0.45/M

Berget.AI $0.46/M

AI/ML API $0.476/M

Vercel AI Gateway $0.5/M

Charm Hyper $0.5/M

Eden AI $0.5/M

Ofox $0.5/M

Regolo AI $0.58/M

Opper $0.581/M

Scaleway $0.684/M

Thinking Machines $0.698/M

Groq $0.8/M

Tempr Gateway $0.8/M

Cerebras $0.99/M

Tempr Gateway $0.99/M

Prime Inference $0.99/M

Cost calculator

Estimate a workload

Provider Input tokens 1,000,000 Output tokens 250,000 Cached input 0%

$0.00

Excludes taxes, non-token charges, and tiered discounts.

Context limits also differ by provider, from 32.768K to 1.13107M tokens. Compare the provider table above before choosing on price alone.

Specification

Capabilities

Recorded from the source catalog and provider listings.

✓ Reasoning Yes

✓ Tool calling Yes

✓ Structured output Yes

✓ Attachments Yes

✓ Vision input Yes

✓ Open weights Yes

Creator: Alibaba Qwen
Model family: qwen
Knowledge cutoff: Not documented
License: Not documented
Release date: 2026-08-14
Model ID: alibaba/qwen3.8-27b

Weights: Hugging Face ↗

Published evaluations

Benchmarks

Every result stays attached to its source, version, metric and harness. Scores from different versions are never merged, and the profile below plots each benchmark against its own population rather than on a shared scale.

Benchmark registry

Benchmark profile

Each axis is this model's percentile among the 8 benchmarks it has published results for, measured against every other model with a score on that same benchmark. Percentiles are used because benchmarks do not share a scale — a 60 on one is not a 60 on another. Hover any point for the raw score.

Agents' Last Exam score Source ↗

Agents' Last Exam #7 of 14

42.9

Humanity's Last Exam score Source ↗

Humanity's Last Exam #12 of 12

30.8

GPQA Diamond score Source ↗

GPQA Diamond #7 of 8

89.2

JobBench score Source ↗

JobBench #7 of 8

33.4

CharXiv Reasoning score Source ↗

CharXiv Reasoning #2 of 5

90.2

IFBench score Source ↗

IFBench #5 of 5

79.5

NL2Repo score · Claude Code Source ↗

NL2Repo #5 of 5

42.3

OSWorld-Verified score Source ↗

OSWorld-Verified #1 of 4

84.3

BabyVision score Source ↗

BabyVision #2 of 3

85.6

CoWorkBench score Source ↗

CoWorkBench #3 of 3

70.7

AndroidWorld score Source ↗

81.9 2 models scored — too few for a distribution

ClawEval-MM average score Source ↗

56.9 2 models scored — too few for a distribution

DeepSWE — 1.1 score · Claude Code Source ↗

42.2 2 models scored — too few for a distribution

ERQA score Source ↗

65.5 2 models scored — too few for a distribution

LiveCodeBench — 6 score Source ↗

90.3 2 models scored — too few for a distribution

MathVision score Source ↗

94.6 2 models scored — too few for a distribution

QwenSWEBench avg@3 · Claude Code Source ↗

79.0 2 models scored — too few for a distribution

RealWorldQA score Source ↗

85.9 2 models scored — too few for a distribution

RecreationBench score Source ↗

47.1 2 models scored — too few for a distribution

Vision2Web score · Claude Code Source ↗

62.9 2 models scored — too few for a distribution

OmniDocBench — 1.5 score Source ↗

91.1 1 model scored — too few for a distribution

SWE-MM score Source ↗

38.6 1 model scored — too few for a distribution

SWE-bench Pro resolved Source ↗

61.7 1 model scored — too few for a distribution

Terminal-Bench — 2.1 score · Terminus Source ↗

73.0 1 model scored — too few for a distribution

WebArena-Verified score · OSWorld scaffold; official WebArena-Verified grader Source ↗

64.8 1 model scored — too few for a distribution

Each strip shows every published score for that benchmark, with this model marked. The lighter shape behind the ticks is the density of results, and the dashed line is the median.

Cost against capability

Price and performance

Listed input price plotted against Agents' Last Exam, the benchmark with the widest published coverage that this model appears in.

The stepped line is the efficient frontier: at each price, the best score available for that money or less. A model sitting on it is not being beaten by anything cheaper. Price is log-scaled because listed rates span four orders of magnitude. Only models with both a listed price and a score on this benchmark can appear.

Catalog activity

Change log

Field-level changes detected between successful source imports.

Full change log

Oct 11, 2026Price Completion0.32 → 1.35

Oct 11, 2026Price Prompt0.045 → 0.04

Oct 02, 2026Price Completion0.4 → 0.32

Oct 02, 2026Price Prompt0.1 → 0.045

Sep 30, 2026Price Completion4.35 → 0.4

Sep 30, 2026Price Prompt0.0249 → 0.1

Sep 30, 2026Price Completion0.4 → 4.35

Sep 30, 2026Price Prompt0.1 → 0.0249

Sep 30, 2026Price Completion4.4 → 0.4

Sep 30, 2026Price Prompt0.0249 → 0.1

Provenance

Sources & verification

Every figure on this page traces back to one of these records.

Methodology

Last verified: Oct 24, 2026
Status: Source-linked
Confidence: Medium
Catalog source: Models.dev

Public API

Use this record

Fetch the complete source-linked model record. No key, no account, no rate-limited tier.

API documentation

Endpoint

GET https://modelbench.lol/api/v1/models/alibaba/qwen3.8-27b

curl

curl "https://modelbench.lol/api/v1/models/alibaba/qwen3.8-27b"

Common questions

Frequently asked questions

Answered directly from the stored record — nothing here is generated beyond the catalog's own fields.

What is Qwen3.8 27B?

Dense 27B vision-language model for coding, agent tasks, and image and video understanding. It is published by Alibaba Qwen and catalogued here from Models.dev.

How much does Qwen3.8 27B cost?

Listed input pricing starts at $0.04 per million tokens from NEAR AI Cloud, rising to $0.99 across 48 listed providers.

What is the context length of Qwen3.8 27B?

Qwen3.8 27B accepts up to 262.144K tokens of context and returns up to 32.768K output tokens.

Does Qwen3.8 27B support tool calling and structured output?

Provider catalogs list support for tool calling, structured output, reasoning, and image input.

Which providers serve Qwen3.8 27B?

44 providers list this model: AMD, IteraCompute, Blue Claw, NEAR AI Cloud, engy, DevPass (LLM Gateway) and 38 more.

Are the weights for Qwen3.8 27B open?

Yes. The weights are published and downloadable from Hugging Face.

When was Qwen3.8 27B released?

The catalog records a release date of 2026-08-14, last verified Oct 24, 2026.

More models from Alibaba Qwen

View all →

Qwen: Qwen3.7 Plus

Qwen3.7-Plus is a cost-effective model in Alibaba's Qwen3.7 series. It supports text and image input with text output, building on the series' text capabilities with a comprehensive upgrade to its...

1M ctx $0.32/M in Qwen: Qwen3.7 Max

Qwen3.7-Max is the flagship model in Alibaba's Qwen3.7 series. It supports text input and output and is designed for agent-centric workloads, with particular strengths in coding, office and productivity tasks,...

1M ctx $1.475/M in Qwen: Qwen3.5 Plus 2026-04-20

Qwen3.5 Plus (April 2026) is a large-scale multimodal language model from Alibaba. It accepts text, image, and video input and produces text output, with a 1M token context window. This...

1M ctx $0.3/M in Qwen: Qwen3.6 Flash

Qwen3.6 Flash is a fast, efficient language model from Alibaba's Qwen 3.6 series. It supports text, image, and video input with a 1M token context window. Tiered pricing kicks in...

1M ctx $0.188/M in