Noometry

Alibaba (Qwen), open weights

Qwen3.8 27B

Qwen3.8 27B by Alibaba (Qwen) ranks 68th of 354 ranked models on the Noometry Index as of October 2026, with a score of 46.0. Its strongest category is multimodal, where it ranks 37th. API pricing starts at $0.99 per million input tokens and $1.49 per million output tokens, with a 262K-token context window.

Last verified

Noometry rank
#68 of 354
Index score
46.0
Evidence
Confirmed 31 results
Provider
Alibaba (Qwen)
Released
August 14, 2026
Weights
Open weights
Reasoning
Yes
Context window
262K
Max output
33K
Input price
$0.99 / M
Output price
$1.49 / M
Blended price
$1.11 / M
Output speed
Not measured
Value
#114 of 219
Knowledge cutoff
Unknown
Input
text, image, video
Hugging Face
Qwen/Qwen3.8-27B

Category scores

Each category score combines every public result we have in that category.

  1. Coding 50.5
  2. Agentic & Tool Use 32.9
  3. Reasoning 41.0
  4. Math 37.1
  5. Knowledge 41.6
  6. Multimodal 41.3
  7. Multilingual 53.7
  8. Instruction Following 75.8
  9. Long Context 44.3
  10. Writing & Preference 65.8
CategoryScoreRankResults
Coding50.5#443
Agentic & Tool Use32.9#571
Reasoning41.0#548
Math37.1#1612
Knowledge41.6#1091
Multimodal41.3#371
Multilingual53.7#601
Instruction Following75.8#531
Long Context44.3#701
Writing & Preference65.8#434

Strengths and weaknesses

Categories where Qwen3.8 27B places highest and lowest among the models ranked in each, with its score against that category's median.

Strongest categories

CategoryScorevs medianRank
Coding50.5+11.8#44 of 340, top 13%
Writing & Preference65.8+12.0#43 of 312, top 14%
Reasoning41.0+17.4#54 of 350, top 16%

Weakest categories

CategoryScorevs medianRank
Math37.1+0.5#161 of 327, top 50%
Agentic & Tool Use32.9+2.5#57 of 154, top 38%
Knowledge41.6+4.3#109 of 314, top 35%

Closest competitors

The models ranked just above and below Qwen3.8 27B. When scores are this close, price and speed are often the better way to choose.

ModelRankScoreBlended $/MSpeed
GPT-5 Pro#6446.4$41.255Compare
Grok 4.20 Multi-Agent#6546.2$1.56—Compare
GLM-5#6646.1$1.5523Compare
Qwen3.5 397B-A17B#6746.0$1.359Compare
GPT-5.3 Codex#6945.8$4.81—Compare
Kimi K2 Thinking Turbo#7045.8——Compare
Qwen3.5 Max Preview#7145.3——Compare
Qwen3.7 Plus#7245.3$0.70—Compare

Sponsored placements are available on pages like this one. Advertise on Noometry

Benchmark results

Every published result we track, with its source. Bold rows are the ones used for ranking; where several exist we prefer independent runs over self-reported numbers.

Coding

BenchmarkScorePositionSettingSourceDate
LMArena WebDev1593#23 of 113, top 21%LMArena2026-10-08
SciCode39.8%lowEpoch AI
SciCode38.1%mediumEpoch AI
SciCode35.6%noneEpoch AI
SciCode46.6%#57 of 121, top 48%xhighEpoch AI
LMArena Coding1482#44 of 294, top 15%LMArena2026-10-08

Agentic & Tool Use

BenchmarkScorePositionSettingSourceDate
APEX-Agents47.5%#28 of 49, top 58%Epoch AI

Reasoning

BenchmarkScorePositionSettingSourceDate
ARC-AGI-222.8%lowEpoch AI
ARC-AGI-213.2%mediumEpoch AI
ARC-AGI-21.5%noneEpoch AI
ARC-AGI-242.4%#35 of 83, top 43%xhighEpoch AI
NYT Connections (extended)54.5%#60 of 91, top 66%Lech Mazur benchmarks
ARC-AGI-169.2%lowEpoch AI
ARC-AGI-168.7%mediumEpoch AI
ARC-AGI-134%noneEpoch AI
ARC-AGI-187.5%#31 of 83, top 38%xhighEpoch AI
CritPt0%lowEpoch AI
CritPt0%mediumEpoch AI
CritPt0.3%noneEpoch AI
CritPt5.4%#57 of 134, top 43%xhighEpoch AI
LMArena Hard Prompts1460#48 of 297, top 17%LMArena2026-10-08
DTBench88%#46 of 151, top 31%Epoch AI
LMCA41.4%#43 of 125, top 35%Epoch AI
Surface Evolver Bench45%#18 of 25, top 72%xhighEpoch AI
Epoch Capabilities Index149.38#57 of 213, top 27%Epoch AI2026-08-14

Math

BenchmarkScorePositionSettingSourceDate
ProofBench16%#56 of 77, top 73%Epoch AI
LMArena Math1456#56 of 285, top 20%LMArena2026-10-08

Knowledge

BenchmarkScorePositionSettingSourceDate
LMArena Expert1482#39 of 273, top 15%LMArena2026-10-08

Multimodal

BenchmarkScorePositionSettingSourceDate
LMArena Vision1271#35 of 122, top 29%LMArena2026-10-09

Multilingual

BenchmarkScorePositionSettingSourceDate
LMArena Non-English1430#60 of 297, top 21%LMArena2026-10-08
LMArena Chinese1504#43 of 285, top 16%LMArena2026-10-08
LMArena French1465#45 of 223, top 21%LMArena2026-10-08
LMArena German1438#59 of 231, top 26%LMArena2026-10-08
LMArena Japanese1384#74 of 211, top 36%LMArena2026-10-08
LMArena Korean1393#63 of 213, top 30%LMArena2026-10-08
LMArena Russian1415#88 of 283, top 32%LMArena2026-10-08
LMArena Spanish1448#54 of 226, top 24%LMArena2026-10-08

Instruction Following

BenchmarkScorePositionSettingSourceDate
LMArena Instruction Following1439#50 of 298, top 17%LMArena2026-10-08

Long Context

BenchmarkScorePositionSettingSourceDate
LMArena Longer Query1450#53 of 291, top 19%LMArena2026-10-08

Writing & Preference

BenchmarkScorePositionSettingSourceDate
LMArena Text1441#60 of 297, top 21%LMArena2026-10-08
LMArena Creative Writing1384#101 of 295, top 35%LMArena2026-10-08
EQ-Bench Creative Writing1671#36 of 115, top 32%EQ-Bench
LMArena Multi-Turn1441#73 of 295, top 25%LMArena2026-10-08

API pricing by provider

RouteInput $/MOutput $/MCached input $/MChecked
cerebras$0.99$1.49$0.992026-10-10
deepinfra$0.20$2.50$0.052026-10-10
groq$0.80$4—2026-10-10
openrouter$0.42$2.55$0.0852026-10-10

All Alibaba (Qwen) API prices → Estimate your cost →

Compare Qwen3.8 27B

Other Alibaba (Qwen) models

Frequently asked questions

How good is Qwen3.8 27B?

Qwen3.8 27B by Alibaba (Qwen) ranks 68th of 354 ranked models on the Noometry Index as of October 2026, with a score of 46.0. Its strongest category is multimodal, where it ranks 37th. API pricing starts at $0.99 per million input tokens and $1.49 per million output tokens, with a 262K-token context window.

How much does Qwen3.8 27B cost?

Qwen3.8 27B costs $0.99 per million input tokens and $1.49 per million output tokens on cerebras, with cached input at $0.99.

What is Qwen3.8 27B's context window?

Qwen3.8 27B accepts up to 262K tokens of input and can write up to 33K tokens in one response.

Is Qwen3.8 27B open source?

Yes. Qwen3.8 27B's weights are downloadable from Hugging Face (Qwen/Qwen3.8-27B); check the license for commercial terms.

What are Qwen3.8 27B's strengths and weaknesses?

Relative to other ranked models, Qwen3.8 27B places best in coding, writing & preference, reasoning and lowest in math, agentic & tool use, knowledge.

What is Qwen3.8 27B best at?

Its best category is multimodal, where it ranks 37th on Noometry.