vs
2:9
overall capability wins (2 ties across 13 scored). Qwen3.8 Flash Next leads overall.
- Reasoning1:5
- Coding0:3
- Agents0:1
- Tool use1:0
- Composite indices1:1
Verdict
Qwen3.8 Flash Next leads coding (widest gap: +42.0 on WebDev Elo).
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Qwen3.8 27B
- -
- Qwen3.8 Flash Next
- $0.27
Price per 1M tokens
Full comparison
Organization
Alibaba
Alibaba
Family
Qwen
Qwen
License
Apache 2.0
Qwen Community License 1.0
Open weights
Yes
Yes
Release
Aug 14, 2026
Aug 26, 2026
Knowledge cutoff
-
-
API / provider
No primary API price listed
Alibaba (QwenCloud)
Modalities
text, image, video → text
text, image, video → text
Specs
Context window
262K
262K
Max output
131K
131K
Parameters
27B
—
Pricing
Input $/1M
—
$0.15
Output $/1M
—
$0.47
Blended $/1M (3∶1)
-
$0.23
Speed
tok/s
-
77.3
TTFT (s)
-
2.79
Reasoning
MMLU-Pro
—
—
GPQA Diamond
90.5
92.3
Humanity's Last Exam
33.9
38.0
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
15.6
24.5
AA-LCR v1.1
82.0
79.7
CritPt
5.4
11.1
MMMU-Pro
76.3
79.8
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
86.1
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
46.6
50.6
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
42.3
—
DeepSWE
42.2
—
WebDev Arena
1,593
1,635
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,437
—
| Metric | Qwen3.8 27B | Qwen3.8 Flash Next | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Alibaba | Alibaba | — |
| Family | Qwen | Qwen | — |
| License | Apache 2.0 | Qwen Community License 1.0 | — |
| Open weights | Yes | Yes | — |
| Release | Aug 14, 2026 | Aug 26, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | No primary API price listed | Alibaba (QwenCloud) | — |
| Modalities | text, image, video → text | text, image, video → text | — |
| Specs | |||
| Context window | 262K | 262K | tie |
| Max output | 131K | 131K | tie |
| Parameters | 27B | — | — |
| Pricing | |||
| Input $/1M | — | $0.15 | — |
| Output $/1M | — | $0.47 | — |
| Blended $/1M (3∶1) | - | $0.23 | — |
| Speed | |||
| tok/s | - | 77.3 | — |
| TTFT (s) | - | 2.79 | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 90.5 | 92.3 | B +1.8 pts |
| Humanity's Last Exam | 33.9 | 38.0 | B +4.1 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 15.6 | 24.5 | B +8.9 pts |
| AA-LCR v1.1 | 82.0 | 79.7 | A +2.3 pts |
| CritPt | 5.4 | 11.1 | B +5.7 pts |
| MMMU-Pro | 76.3 | 79.8 | B +3.5 pts |
| 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 | 86.1 | B +6.3 pts |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 46.6 | 50.6 | B +4.0 pts |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | 42.3 | — | — |
| DeepSWE | 42.2 | — | — |
| WebDev Arena | 1,593 | 1,635 | B +42 Elo |
| 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).
Qwen3.8 Flash Next
- Context
- 262K / 131K out
- Parameters
- —
- Price
- $0.15 / $0.47
- Speed
- 77.3 tok/s · 2.79s TTFT
- Modalities
- text, image, video → text
- License
- Qwen Community License 1.0
Open-weight sparse-MoE Flash tier (6B active) previewing the Qwen4 architecture, released Aug 26, 2026 — 262K native context extensible to ~1M with YaRN, tuned for cost efficiency at $0.15/$0.47 per 1M (cache-hit $0.016). Reported total parameter count varies by source (~125-180B). Benchmark figures are Alibaba's self-reported launch numbers (SWE-bench Pro 62.5, HLE 35.9); Artificial Analysis places it at 56 on its Intelligence Index.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.