Ling 3.1 Flash

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-09
RankedEnough results for an index.
IndexAll categories, out of 100.65.9±7.2Median 50.7
CoverageIndex weight that has results.70%
SpeedOutput tokens per second.78/sMedian 76/s
First tokenSeconds to the first output token.1.1 sMedian 1.1 s
Input / 1MUS dollars per 1M input tokens.$0.06Median $0.60
Output / 1MUS dollars per 1M output tokens.$0.18Median $2.50
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1425

Details

From OpenRouter
API IDThe model ID on OpenRouter.
inclusionai/ling-3.1-flash
Max outputMaximum output tokens per request.
262K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
No
Cache read / 1MUS dollars per 1M cached tokens.
$0.012
Weights
Hugging Face (opens in a new tab) ↗
RetiresDate OpenRouter removes the model.
—

Run locally

572.12B parameters
16-bit
1202.5 GB
8-bit
639.3 GB
4-bit
365.2 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory for a short context. 4-bit needs more than 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.9 counted65.445 / 221
CodingCode writing and repair.2 counted65.736 / 214
ReasoningLogic problems and puzzles.2 counted66.734 / 225
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.2 counted69.025 / 266
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.None counted——

Results

15 counted
BenchmarkResultThe score from the source.PlacePlace among results on this benchmark.
BrowseCompAgentic91.7%5 / 63
SWE-bench ProCoding68.7%10 / 83
GDPval-AA normalizedAgentic55.9%16 / 132
WideResearchAgentic83.3%2 / 16
CyberGymAgentic87.9%4 / 29
Critical Physics TasksReasoning18.0%34 / 207
MCP AtlasAgentic83.0%11 / 53
HealthBench ProfessionalKnowledge65.3%5 / 21
AutomationBenchAgentic52.5%8 / 33
Artificial Analysis SciCodeCoding54.1%36 / 113
Humanity's Last ExamKnowledge37.4%39 / 77
DeepSWEAgentic59.7%46 / 64

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reusemodels.devMITCheaper InferenceNo licence stated

History

13 releases
Index67.5 → 65.9
23 Oct 202621 Nov 2026
Output / 1M$0.225 → $0.18
23 Oct 202621 Nov 2026

Questions

How much does Ling 3.1 Flash cost?
$0.06 per 1M input tokens, $0.18 per 1M output tokens.
How good is Ling 3.1 Flash?
Its index is 65.9 out of 100. That is place 39 of 282 ranked models.
What is Ling 3.1 Flash best at?
Knowledge: 69.0 out of 100, place 25 of 266.
How fast is Ling 3.1 Flash?
78 output tokens per second, 1.1 s to the first token.
What is the context window of Ling 3.1 Flash?
1M tokens.
Is Ling 3.1 Flash open source?
Yes. Its weights are published.
When was Ling 3.1 Flash released?
September 2026.

Badge

Ling 3.1 Flash on Inference360: place 39 of 282, index 65.9

Compare with

Kimi K365.9Hy4 preview66.2Inkling 2 Small65.5Gemini 3.8 Flash66.5Hy4 flash64.5

Lower price

Hy4 flash64.5 · $0.14/1M

More from InclusionAI

All models
Ming-3.0-asr-tinyUnrankedMing-3.0-audio-flashUnrankedMing-3.0-audio-tinyUnrankedLing 3.0 Flash SanteUnrankedLing 3.0 Flash VLUnranked