Inference
EN

Qwen3.5 Plus Thinking vs GPT-5 mini

Change models
Qwen3.5 Plus ThinkingAlibaba
GPT-5 miniOpenAI
IndexAll categories, out of 100.
48.1 ±8.2
48.4 ±3.9 best
GapIndex points below the top.
−0.3 · no clear gap
Top
CoverageIndex weight that has results.
75%
95% best
SpeedOutput tokens per second.
—
78/s
EloLMArena rating and rank.
—
1373
ReleasedMonth of release.
—
2025-08
Input / 1MUS dollars per 1M input tokens.
—
$0.25
Output / 1MUS dollars per 1M output tokens.
—
$2.00
Batch output / 1MBatch price per 1M output tokens.
—
$1.00
1K chats2K tokens in, 500 out, each.
—
$1.50
1K documents20K tokens in, 1K out, each.
—
$7.00
1K agent tasks200K tokens in, 20K out, each.
—
$90.00
ContextMaximum tokens in one request.
—
400K
AgenticMulti-step tasks with tools.
51.6 best
41.4
CodingCode writing and repair.
53.9 best
49.4
ReasoningLogic problems and puzzles.
—
47.0
MultimodalTasks with images and text.
5.0
57.9 best
KnowledgeFacts and expert knowledge.
55.8 best
46.1
MultilingualTasks in many languages.
—
51.8
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
53.1 best
52.2

Benchmarks

14 shared
Only GPT-5 mini: 36
Qwen3.5 Plus Thinking
GPT-5 mini
LegalBenchLegal
85.1% best25 / 149
81.8%80 / 149
LiveCodeBenchCoding
85.3%33 / 142
86.6% best21 / 142
GPQA Diamond · Vals AIKnowledge
87.4% best38 / 135
80.3%66 / 135
MMLU ProKnowledge
87.2% best32 / 135
82.2%78 / 135
CorpFin v2Finance
65.3% best28 / 131
60.2%70 / 131
Vibe Code Bench v1.1Coding
15.7% best97 / 123
14.2%100 / 123
MortgageTaxFinance
60.8%64 / 97
66.9% best31 / 97
AIMEMath
86.0%33 / 95
91.5% best24 / 95
MedQAHealthcare
95.2%13 / 95
96.1% best5 / 95
MMMU ProMultimodal
22.8%92 / 92
78.9% best50 / 92
SWE-benchCoding
71.2% best59 / 87
60.8%76 / 87
Terminal-Bench 2.0 · Vals AIAgentic
41.6% best29 / 66
27.0%50 / 66
CaseLaw v2Legal
59.7%25 / 53
68.5% best4 / 53
Finance Agent v1.1Finance
54.5% best16 / 51
51.9%26 / 51