vs
0:3
overall capability wins (0 ties across 3 scored). Qwen3.8 Max leads overall.
- Pricing3:0
- Speed1:1
- Specs0:3
Verdict
Qwen3.8 Max offers ~7.6x larger context (1M vs 131K); Qwen3.8 Max allows ~16.0x more max output (131K vs 8K); Qwen2.5 32B Instruct is ~15.0x cheaper on a blended token basis than Qwen3.8 Max.
- Qwen3.8 Max offers ~7.6x larger context (1M vs 131K)
- Qwen3.8 Max allows ~16.0x more max output (131K vs 8K)
- Qwen2.5 32B Instruct is ~15.0x cheaper on a blended token basis than Qwen3.8 Max
Cost for 1M in + 250K out
Illustrative chat workload at primary-provider list prices.
- Qwen2.5 32B Instruct
- $0.25
- Qwen3.8 Max
- $3.50
Qwen2.5 32B Instruct is about 14.0x cheaper on this workload.
Price per 1M tokens
Full comparison
Organization
Alibaba
Alibaba
Family
Qwen
Qwen
License
Apache 2.0
Qwen3.8-Max License
Open weights
Yes
Yes
Release
Sep 19, 2024
Aug 3, 2026
Knowledge cutoff
-
-
API / provider
Together / Fireworks (ref.)
Alibaba Cloud
Modalities
text → text
text, image, video → text
Specs
Context window
131K
1M
Max output
8K
131K
Parameters
32B
2.4T (95B act.)
Pricing
Input $/1M
$0.20
$2
Output $/1M
$0.20
$6
Blended $/1M (3∶1)
$0.20
$3
Speed
tok/s
80
86
TTFT (s)
0.3
0.4
Reasoning
MMLU-Pro
69.0
—
GPQA Diamond
—
92.8
Humanity's Last Exam
—
43.1
AIME 2025
—
—
MATH-500
—
—
Humanity's Last Exam (with tools)
—
—
AA-Omniscience Accuracy
—
31.7
AA-LCR v1.1
—
80.3
CritPt
—
17.7
MMMU-Pro
—
82.8
IFBench
—
—
Chartography
—
—
Chartography (With Tools)
—
—
Coding
SWE-bench Verified
—
—
SWE-bench Pro
—
67.7
SWE-bench Multilingual
—
—
LiveCodeBench
51.2
—
Terminal-Bench 2.1
—
81.3
Aider Polyglot
—
—
Terminal-Bench 3
—
—
BigCodeBench
45.0
—
SciCode
—
52.1
CursorBench
—
—
SWE-Rebench
—
—
NL2Repo-Bench
—
55.9
DeepSWE
—
56.6
WebDev Arena
—
1,671
Terminal-Bench 4.0
—
—
CursorBench 4.0
—
—
FrontierCode v1.1 (Main)
—
—
Arena
LMArena Elo
—
1,481
| Metric | Qwen2.5 32B Instruct | Qwen3.8 Max | Delta |
|---|---|---|---|
| Identity | |||
| Organization | Alibaba | Alibaba | — |
| Family | Qwen | Qwen | — |
| License | Apache 2.0 | Qwen3.8-Max License | — |
| Open weights | Yes | Yes | — |
| Release | Sep 19, 2024 | Aug 3, 2026 | — |
| Knowledge cutoff | - | - | — |
| API / provider | Together / Fireworks (ref.) | Alibaba Cloud | — |
| Modalities | text → text | text, image, video → text | — |
| Specs | |||
| Context window | 131K | 1M | B +869K |
| Max output | 8K | 131K | B +123K |
| Parameters | 32B | 2.4T (95B act.) | B +2K |
| Pricing | |||
| Input $/1M | $0.20 | $2 | A +$1.80 |
| Output $/1M | $0.20 | $6 | A +$5.80 |
| Blended $/1M (3∶1) | $0.20 | $3 | A +$2.80 |
| Speed | |||
| tok/s | 80 | 86 | B +6.00 |
| TTFT (s) | 0.3 | 0.4 | A +0.10 |
| Reasoning | |||
| MMLU-Pro | 69.0 | — | — |
| GPQA Diamond | — | 92.8 | — |
| Humanity's Last Exam | — | 43.1 | — |
| AIME 2025 | — | — | — |
| MATH-500 | — | — | — |
| Humanity's Last Exam (with tools) | — | — | — |
| AA-Omniscience Accuracy | — | 31.7 | — |
| AA-LCR v1.1 | — | 80.3 | — |
| CritPt | — | 17.7 | — |
| MMMU-Pro | — | 82.8 | — |
| IFBench | — | — | — |
| Chartography | — | — | — |
| Chartography (With Tools) | — | — | — |
| Coding | |||
| SWE-bench Verified | — | — | — |
| SWE-bench Pro | — | 67.7 | — |
| SWE-bench Multilingual | — | — | — |
| LiveCodeBench | 51.2 | — | — |
| Terminal-Bench 2.1 | — | 81.3 | — |
| Aider Polyglot | — | — | — |
| Terminal-Bench 3 | — | — | — |
| BigCodeBench | 45.0 | — | — |
| SciCode | — | 52.1 | — |
| CursorBench | — | — | — |
| SWE-Rebench | — | — | — |
| NL2Repo-Bench | — | 55.9 | — |
| DeepSWE | — | 56.6 | — |
| WebDev Arena | — | 1,671 | — |
| Terminal-Bench 4.0 | — | — | — |
| CursorBench 4.0 | — | — | — |
| FrontierCode v1.1 (Main) | — | — | — |
| Arena | |||
| LMArena Elo | — | 1,481 | — |
Qwen2.5 32B Instruct
- Context
- 131K / 8K out
- Parameters
- 32B
- Price
- $0.20 / $0.20
- Speed
- 80 tok/s · 0.3s TTFT
- Modalities
- text → text
- License
- Apache 2.0
Mid-size Qwen2.5 instruct — strong open multilingual chat between the 14B and 72B tiers.
Qwen3.8 Max
- Context
- 1M / 131K out
- Parameters
- 2.4T (95B act.)
- Price
- $2 / $6
- Speed
- 86 tok/s · 0.4s TTFT
- Modalities
- text, image, video → text
- License
- Qwen3.8-Max License
Alibaba's Aug 2026 Max flagship (2.4T MoE / 95B active) — frontier agentic coding via QwenCloud / DashScope with 1M multimodal context and $2/$6 API pricing. Open weights released Aug 12, 2026 as Qwen3.8-2.4T-A95B (text-only, thinking-mandatory, Qwen3.8-Max license); the hosted version adds vision input, non-thinking mode, and built-in tools.
Benchmark charts
Winner bars are emphasized. Per-benchmark deltas sit above each chart.
These two models have no overlapping published benchmarks in our dataset. Compare specs and pricing instead.