vs
7:2
overall capability wins (0 ties across 9 scored). Qwen3.8 27B leads overall.
- Reasoning3:1
- Coding2:0
- Specs2:1
Verdict
Qwen3.8 27B leads coding (widest gap: +50.0 on LiveCode); Qwen3.8 27B offers ~1.6x larger context (262K vs 160K); Qwen3.8 27B allows ~8.0x more max output (131K vs 16K).
- Qwen3.8 27B leads coding (widest gap: +50.0 on LiveCode)
- Qwen3.8 27B offers ~1.6x larger context (262K vs 160K)
- Qwen3.8 27B allows ~8.0x more max output (131K vs 16K)
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Qwen3.8 27B
- -
- Qwen3-Coder 30B-A3B
- $0.14
Price per 1M tokens
Full comparison
Organization
Alibaba
Alibaba
Family
Qwen
Qwen
License
Apache 2.0
Apache 2.0
Open weights
Yes
Yes
Release
Aug 14, 2026
Jul 22, 2025
Knowledge cutoff
-
-
API / provider
No primary API price listed
Alibaba / Together (ref.)
Modalities
text, image, video → text
text → text
Specs
Context window
262K
160K
Max output
131K
16K
Parameters
27B
30B (3B act.)
Pricing
Input $/1M
—
$0.07
Output $/1M
—
$0.27
Blended $/1M (3∶1)
-
$0.12
Speed
tok/s
-
120
TTFT (s)
-
0.25
Reasoning
MMLU-Pro
—
70.6
GPQA Diamond
90.5
51.6
Humanity's Last Exam
33.9
3.8
AIME 2025
—
29.0
MATH-500
—
89.3
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
15.6
16.4
AA-LCR v1.1
82.0
32.7
CritPt
5.4
—
MMMU-Pro
76.3
—
IFBench
—
32.7
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
61.7
—
SWE-bench Multilingual
—
—
LiveCodeBench
90.3
40.3
Terminal-Bench 2.1
79.8
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
46.6
27.8
CursorBench
—
—
SWE-Rebench
—
21.6
NL2Repo-Bench
42.3
—
DeepSWE
42.2
—
WebDev Arena
1,593
—
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,437
—
| Metric | Qwen3.8 27B | Qwen3-Coder 30B-A3B | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Alibaba | Alibaba | — |
| Family | Qwen | Qwen | — |
| License | Apache 2.0 | Apache 2.0 | — |
| Open weights | Yes | Yes | — |
| Release | Aug 14, 2026 | Jul 22, 2025 | — |
| Knowledge cutoff | - | - | — |
| API / provider | No primary API price listed | Alibaba / Together (ref.) | — |
| Modalities | text, image, video → text | text → text | — |
| Specs | |||
| Context window | 262K | 160K | A +102K |
| Max output | 131K | 16K | A +115K |
| Parameters | 27B | 30B (3B act.) | B +3.00 |
| Pricing | |||
| Input $/1M | — | $0.07 | — |
| Output $/1M | — | $0.27 | — |
| Blended $/1M (3∶1) | - | $0.12 | — |
| Speed | |||
| tok/s | - | 120 | — |
| TTFT (s) | - | 0.25 | — |
| Reasoning | |||
| MMLU-Pro | — | 70.6 | — |
| GPQA Diamond | 90.5 | 51.6 | A +38.9 pts |
| Humanity's Last Exam | 33.9 | 3.8 | A +30.1 pts |
| AIME 2025 | — | 29.0 | — |
| MATH-500 | — | 89.3 | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 15.6 | 16.4 | B +0.8 pts |
| AA-LCR v1.1 | 82.0 | 32.7 | A +49.3 pts |
| CritPt | 5.4 | — | — |
| MMMU-Pro | 76.3 | — | — |
| IFBench | — | 32.7 | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | 61.7 | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 90.3 | 40.3 | A +50.0 pts |
| Terminal-Bench 2.1 | 79.8 | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 46.6 | 27.8 | A +18.8 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | 21.6 | — |
| NL2Repo-Bench | 42.3 | — | — |
| DeepSWE | 42.2 | — | — |
| WebDev Arena | 1,593 | — | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,437 | — | — |
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).
Qwen3-Coder 30B-A3B
- Context
- 160K / 16K out
- Parameters
- 30B (3B act.)
- Price
- $0.07 / $0.27
- Speed
- 120 tok/s · 0.25s TTFT
- Modalities
- text → text
- License
- Apache 2.0
Sparse coding MoE in the Qwen3-Coder line — efficient agent edits without full 480B cost.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.