Inference
EN
Alibaba

Qwen3.5-122B-A10B

Qwen3.5-122B-A10B is a reasoning model from Alibaba. 24 benchmarks count toward its index, in 7 categories.

  • Open weights
  • Reasoning
  • Input: text · image · video
  • Released 2026-03
RankedEnough results for an index.
IndexAll categories, out of 100.51.9±4.3Median 51.0
CoverageIndex weight that has results.90%
SpeedOutput tokens per second.70/sMedian 67/s
First tokenSeconds to the first output token.0.6 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.26Median $0.60
Output / 1MUS dollars per 1M output tokens.$2.08Median $2.55
ContextMaximum tokens in one request.262KMedian 262K
EloLMArena rating and rank.1417±4 #126Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
qwen/qwen3.5-122b-a10b
Max outputMaximum output tokens per request.
66K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
—
WeightsLink to the published weights.
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
5 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
OpenRouterRouter$0.26$2.08262K
SiliconFlowHost$0.26$2.08262K
Deep InfraDeprecatedHost$0.29$2.40262K
Novita AIHost$0.40$3.20262K
RegoloHost$1.12$4.71262K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Run locally

125.09B parameters
16-bit
263.7 GB
8-bit
140.6 GB
4-bit
80.6 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.5 counted51.4120 / 218
CodingCode writing and repair.3 counted50.4135 / 209
ReasoningLogic problems and puzzles.4 counted50.3112 / 227
MultimodalTasks with images and text.5 counted53.958 / 84
KnowledgeFacts and expert knowledge.4 counted52.9115 / 268
MultilingualTasks in many languages.1 counted51.6—
InstructionTasks with strict prompt rules.2 counted55.524 / 63
MathMath problems.None counted——

Results

24 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis IFBenchInstruction75.7%19 / 146
Massive Multitask Language Understanding ProfessionalKnowledge86.7%10 / 75
Instruction-Following EvalInstruction93.4%14 / 70
τ²-Bench Tool-Agent-User EvaluationAgentic93.6%30 / 149
Massive Multi-discipline Multimodal UnderstandingMultimodal83.9%6 / 28
Graduate-Level Google-Proof Q&AKnowledge86.6%46 / 141
Artificial Analysis GPQA DiamondKnowledge85.7%72 / 199
Artificial Analysis Long Context ReasoningReasoning76.3%86 / 217
Artificial Analysis Humanity's Last ExamKnowledge25.2%99 / 222
Artificial Analysis MMMU-ProMultimodal75.0%53 / 113
V*Multimodal93.2%7 / 14
MMLU-ProXMultilingual82.2%9 / 17
LiveCodeBench v6Coding78.9%32 / 59
Software Engineering Benchmark VerifiedCoding72.0%62 / 114
SuperGPQA: Scaling LLM Evaluation Across 285 Graduate DisciplinesKnowledge67.1%12 / 22
Massive Multi-discipline Multimodal Understanding ProMultimodal76.9%30 / 53
Critical Physics TasksReasoning0.6%132 / 214
MathVisionMultimodal86.2%15 / 24
Mystery Game PuzzlesReasoning17.0% ±3.851 / 80
LongBench v2Reasoning60.2%16 / 25
GDPval-AA normalizedAgentic15.7%86 / 133
Terminal-Bench 2.0 · BenchLMAgentic49.4%31 / 41
BrowseCompAgentic63.8%49 / 63
Artificial Analysis SciCodeCoding39.7%89 / 114
OSWorld-VerifiedAgentic58.0%33 / 37
CharXiv ReasoningMultimodal77.2%44 / 49
Artificial Analysis Global-MMLU-LiteMultilingual89.1%—
Multimodal Multi-disciplinary Video UnderstandingMultimodal74.7%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYmodels.devMIT

History

5 releases
Index52.3 → 51.9
23 Oct 202612 Nov 2026
Output / 1M$2.08 → $2.08
23 Oct 202612 Nov 2026

Questions

How much does Qwen3.5-122B-A10B cost?
$0.26 per 1M input tokens, $2.08 per 1M output tokens.
How good is Qwen3.5-122B-A10B?
Its index is 51.9 out of 100. That is place 133 of 283 ranked models.
What is Qwen3.5-122B-A10B best at?
Instruction: 55.5 out of 100, place 24 of 63.
How fast is Qwen3.5-122B-A10B?
70 output tokens per second, 0.6 s to the first token.
What is the context window of Qwen3.5-122B-A10B?
262K tokens.
Is Qwen3.5-122B-A10B open source?
Yes. Its weights are published.
When was Qwen3.5-122B-A10B released?
March 2026.

Badge

Qwen3.5-122B-A10B on Inference360: place 133 of 283, index 51.9
HTML
<a href="https://inference360.ai/models/qwen3-5-122b-a10b"><img src="https://inference360.ai/badges/qwen3-5-122b-a10b.svg" alt="Qwen3.5-122B-A10B on Inference360: place 133 of 283, index 51.9" height="20"></a>
Markdown
[![Qwen3.5-122B-A10B on Inference360: place 133 of 283, index 51.9](https://inference360.ai/badges/qwen3-5-122b-a10b.svg)](https://inference360.ai/models/qwen3-5-122b-a10b)

Compare with

GLM-5-Turbo51.9MiMo-V2-Pro51.8o351.8Qwen3.5-27B51.3Qwen3.5 397B53.8

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Alibaba

All models
Qwen3.5-27B51.3Qwen3.5-35B-A3B49.4Qwen3.5 FlashUnrankedQwen3.5 PlusUnrankedQwen3.5 9BUnranked