vs
0:4
overall capability wins (0 ties across 4 scored). Qwen3.8 Flash Next leads overall.
- Reasoning0:2
- Pricing1:2
- Speed1:1
- Specs0:2
Verdict
Qwen3.8 Flash Next edges reasoning & knowledge; Qwen3.8 Flash Next offers ~8.0x larger context (262K vs 33K); Qwen3.8 Flash Next allows ~16.0x more max output (131K vs 8K).
- Qwen3.8 Flash Next edges reasoning & knowledge
- Qwen3.8 Flash Next offers ~8.0x larger context (262K vs 33K)
- Qwen3.8 Flash Next allows ~16.0x more max output (131K vs 8K)
- Qwen3.8 Flash Next is ~1.6x cheaper on a blended token basis than Qwen2 72B Instruct
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Qwen2 72B Instruct
- $0.45
- Qwen3.8 Flash Next
- $0.27
Qwen3.8 Flash Next is about 1.7x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Alibaba
Alibaba
Family
Qwen
Qwen
License
Tongyi Qianwen
Qwen Community License 1.0
Open weights
Yes
Yes
Release
Jun 7, 2024
Aug 26, 2026
Knowledge cutoff
-
-
API / provider
Together / Fireworks (ref.)
Alibaba (QwenCloud)
Modalities
text → text
text, image, video → text
Specs
Context window
33K
262K
Max output
8K
131K
Parameters
72B
—
Pricing
Input $/1M
$0.35
$0.15
Output $/1M
$0.40
$0.47
Blended $/1M (3∶1)
$0.36
$0.23
Speed
tok/s
65
77.3
TTFT (s)
0.4
2.79
Reasoning
MMLU-Pro
—
—
GPQA Diamond
37.1
92.3
Humanity's Last Exam
3.7
38.0
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
24.5
AA-LCR v1.1
—
79.7
CritPt
—
11.1
MMMU-Pro
—
79.8
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
—
SWE-bench Multilingual
—
—
LiveCodeBench
—
—
Terminal-Bench 2.1
—
86.1
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
—
50.6
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
—
1,635
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,262
—
| Metric | Qwen2 72B Instruct | Qwen3.8 Flash Next | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Alibaba | Alibaba | — |
| Family | Qwen | Qwen | — |
| License | Tongyi Qianwen | Qwen Community License 1.0 | — |
| Open weights | Yes | Yes | — |
| Release | Jun 7, 2024 | Aug 26, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Together / Fireworks (ref.) | Alibaba (QwenCloud) | — |
| Modalities | text → text | text, image, video → text | — |
| Specs | |||
| Context window | 33K | 262K | B +229K |
| Max output | 8K | 131K | B +123K |
| Parameters | 72B | — | — |
| Pricing | |||
| Input $/1M | $0.35 | $0.15 | B +$0.20 |
| Output $/1M | $0.40 | $0.47 | A +$0.07 |
| Blended $/1M (3∶1) | $0.36 | $0.23 | B +$0.13 |
| Speed | |||
| tok/s | 65 | 77.3 | B +12 |
| TTFT (s) | 0.4 | 2.79 | A +2.39 |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 37.1 | 92.3 | B +55.2 pts |
| Humanity's Last Exam | 3.7 | 38.0 | B +34.3 pts |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 24.5 | — |
| AA-LCR v1.1 | — | 79.7 | — |
| CritPt | — | 11.1 | — |
| MMMU-Pro | — | 79.8 | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | — | 86.1 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | — | 50.6 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | — | 1,635 | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,262 | — | — |
Qwen2 72B Instruct
- Context
- 33K / 8K out
- Parameters
- 72B
- Price
- $0.35 / $0.40
- Speed
- 65 tok/s · 0.4s TTFT
- Modalities
- text → text
- License
- Tongyi Qianwen
Prior-generation Qwen2 72B instruct — predecessor to Qwen2.5 still seen in older hosted stacks.
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.