vs
0:2
overall capability wins (1 ties across 3 scored). Qwen3.8 Max 0902 leads overall.
- Coding0:1
- Pricing0:3
- Specs0:1
Verdict
Qwen3.8 Max 0902 leads coding (widest gap: +163.7 on WebDev Elo); Qwen3.8 Max 0902 allows ~2.0x more max output (131K vs 66K); Qwen3.8 Max 0902 is ~1.3x cheaper on a blended token basis than Qwen3.7 Max.
- Qwen3.8 Max 0902 leads coding (widest gap: +163.7 on WebDev Elo)
- Qwen3.8 Max 0902 allows ~2.0x more max output (131K vs 66K)
- Qwen3.8 Max 0902 is ~1.3x cheaper on a blended token basis than Qwen3.7 Max
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Qwen3.7 Max
- $4.38
- Qwen3.8 Max 0902
- $3.50
Qwen3.8 Max 0902 is about 1.3x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Alibaba
Alibaba
Family
Qwen
Qwen
License
Proprietary
Proprietary
Open weights
No
No
Release
May 21, 2026
Sep 2, 2026
Knowledge cutoff
-
-
API / provider
Alibaba Cloud
Alibaba Cloud
Modalities
text, image → text
text, image, video → text
Specs
Context window
1M
1M
Max output
66K
131K
Parameters
—
2.4T (95B act.)
Pricing
Input $/1M
$2.50
$2
Output $/1M
$7.50
$6
Blended $/1M (3∶1)
$3.75
$3
Speed
tok/s
70
-
TTFT (s)
0.45
-
Reasoning
MMLU-Pro
—
—
GPQA Diamond
92.3
—
Humanity's Last Exam
40.5
—
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
31.1
—
AA-LCR v1.1
79.0
—
CritPt
13.4
—
MMMU-Pro
—
—
IFBench
80.5
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
80.4
—
SWE-bench Pro
—
—
SWE-bench Multilingual
78.3
—
LiveCodeBench
—
—
Terminal-Bench 2.1
74.5
—
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
—
—
SciCode
49.5
—
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
—
DeepSWE
—
—
WebDev Arena
1,517
1,681
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
1,473
—
| Metric | Qwen3.7 Max | Qwen3.8 Max 0902 | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Alibaba | Alibaba | — |
| Family | Qwen | Qwen | — |
| License | Proprietary | Proprietary | — |
| Open weights | No | No | — |
| Release | May 21, 2026 | Sep 2, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Alibaba Cloud | Alibaba Cloud | — |
| Modalities | text, image → text | text, image, video → text | — |
| Specs | |||
| Context window | 1M | 1M | tie |
| Max output | 66K | 131K | B +66K |
| Parameters | — | 2.4T (95B act.) | — |
| Pricing | |||
| Input $/1M | $2.50 | $2 | B +$0.50 |
| Output $/1M | $7.50 | $6 | B +$1.50 |
| Blended $/1M (3∶1) | $3.75 | $3 | B +$0.75 |
| Speed | |||
| tok/s | 70 | - | — |
| TTFT (s) | 0.45 | - | — |
| Reasoning | |||
| MMLU-Pro | — | — | — |
| GPQA Diamond | 92.3 | — | — |
| Humanity's Last Exam | 40.5 | — | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | 31.1 | — | — |
| AA-LCR v1.1 | 79.0 | — | — |
| CritPt | 13.4 | — | — |
| MMMU-Pro | — | — | — |
| IFBench | 80.5 | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | 80.4 | — | — |
| SWE-bench Pro | — | — | — |
| SWE-bench Multilingual | 78.3 | — | — |
| LiveCodeBench | — | — | — |
| Terminal-Bench 2.1 | 74.5 | — | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | — | — | — |
| SciCode | 49.5 | — | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | — | — |
| DeepSWE | — | — | — |
| WebDev Arena | 1,517 | 1,681 | B +164 Elo |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | 1,473 | — | — |
Qwen3.7 Max
- Context
- 1M / 66K out
- Parameters
- —
- Price
- $2.50 / $7.50
- Speed
- 70 tok/s · 0.45s TTFT
- Modalities
- text, image → text
- License
- Proprietary
Alibaba's closed Max flagship - near-frontier coding and reasoning via DashScope.
Qwen3.8 Max 0902
- Context
- 1M / 131K out
- Parameters
- 2.4T (95B act.)
- Price
- $2 / $6
- Speed
- —
- Modalities
- text, image, video → text
- License
- Proprietary
September 2, 2026 hosted post-training snapshot of Qwen3.8-Max, further trained for coding and workplace/agent workflows. Same 2.4T MoE / 95B active architecture, 1M multimodal context, and $2/$6 QwenCloud pricing as the August 3 Max release. Arena Code/WebDev listed qwen3.8-max-0902 at 1688 (preliminary) on the Sep 2, 2026 snapshot.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.