vs
1:3
overall capability wins (0 ties across 4 scored). Qwen3.8 27B leads overall.
- Coding0:1
- Arena0:1
- Specs1:1
Verdict
Qwen3.8 27B leads coding (widest gap: +355.1 on WebDev Elo); Qwen3.8 27B ranks higher on LMArena (+40 Elo); Qwen3.5 Flash offers ~3.8x larger context (1M vs 262K).
- Qwen3.8 27B leads coding (widest gap: +355.1 on WebDev Elo)
- Qwen3.8 27B ranks higher on LMArena (+40 Elo)
- Qwen3.5 Flash offers ~3.8x larger context (1M vs 262K)
- Qwen3.8 27B allows ~4.0x more max output (131K vs 33K)
- Qwen3.8 27B ships open weights (Apache 2.0)
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Qwen3.5 Flash
- $0.13
- Qwen3.8 27B
- -
Price per 1M tokens
Full comparison
Organization
Alibaba
Alibaba
Family
Qwen
Qwen
License
Proprietary
Apache 2.0
Open weights
No
Yes
Release
Mar 1, 2026
Aug 14, 2026
Knowledge cutoff
-
-
API / provider
Alibaba
No primary API price listed
Modalities
text, image → text
text, image, video → text
Specs
Context window
1M
262K
Max output
33K
131K
Parameters
—
27B
Pricing
Input $/1M
$0.07
—
Output $/1M
$0.26
—
Blended $/1M (3∶1)
$0.11
-
Speed
tok/s
150
-
TTFT (s)
0.2
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
—
90.5
Humanity's Last Exam
—
33.9
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
15.6
AA-LCR v1.1
—
82.0
CritPt
—
5.4
MMMU-Pro
—
76.3
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
61.7
SWE-bench Multilingual
—
—
LiveCodeBench
—
90.3
Terminal-Bench 2.1
—
79.8
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
46.6
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
42.3
DeepSWE
—
42.2
WebDev Arena
1,238
1,593
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,397
1,437
| Metric | Qwen3.5 Flash | Qwen3.8 27B | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Alibaba | Alibaba | — |
| Family | Qwen | Qwen | — |
| License | Proprietary | Apache 2.0 | — |
| Open weights | No | Yes | — |
| Release | Mar 1, 2026 | Aug 14, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Alibaba | No primary API price listed | — |
| Modalities | text, image → text | text, image, video → text | — |
| Specs | |||
| Context window | 1M | 262K | A +738K |
| Max output | 33K | 131K | B +98K |
| Parameters | — | 27B | — |
| Pricing | |||
| Input $/1M | $0.07 | — | — |
| Output $/1M | $0.26 | — | — |
| Blended $/1M (3∶1) | $0.11 | - | — |
| Speed | |||
| tok/s | 150 | - | — |
| TTFT (s) | 0.2 | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | — | 90.5 | — |
| Humanity's Last Exam | — | 33.9 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 15.6 | — |
| AA-LCR v1.1 | — | 82.0 | — |
| CritPt | — | 5.4 | — |
| MMMU-Pro | — | 76.3 | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | 61.7 | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | 90.3 | — |
| Terminal-Bench 2.1 | — | 79.8 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | 46.6 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | 42.3 | — |
| DeepSWE | — | 42.2 | — |
| WebDev Arena | 1,238 | 1,593 | B +355 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,397 | 1,437 | B +40 Elo |
Qwen3.5 Flash
- Context
- 1M / 33K out
- Parameters
- —
- Price
- $0.07 / $0.26
- Speed
- 150 tok/s · 0.2s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Low-cost Qwen3.5 API Flash tier with million-token context for high-volume Chinese/English chat.
Qwen3.8 27B
- Context
- 262K / 131K out
- Parameters
- 27B
- Price
- —
- Speed
- —
- Modalities
- text, image, video → text
- License
- Apache 2.0
Open-weight Qwen3.8 27B dense vision-language model (Apache-2.0, released Aug 14, 2026) — native image and video understanding, 262K context extensible to 1M, flexible thinking control via reasoning_effort, and strong agentic coding for its size (Terminal-Bench 2.1 73.0, SWE-bench Pro 61.7, GPQA Diamond 89.2).
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.