Data as of November 13, 2026 · How the score is built
Ling 3.1 Flash
Decision readingLing 3.1 Flash scores 64.7 out of 100 and ranks #30 of 215. This profile shows 18 source-displayable benchmark rows; its strongest eligible category is Reasoning at #28. No comparable first-party API price is published in the catalog.
Email me when its price, rank or availability changes.
Capability
64.7/100
field median 49.4#30 of 215 ranked models
Public
#30of 215
Verified —
Price
API rate not published
input median $0.60No comparable first-party hosted token rate
Speed
223tok/s
field median 92 tok/sFirst token 10.72 s
Context
262Ktokens
field median 256,000Maximum output length is tracked separately
Decision snapshot
Where this model sits in the ranked field, its strongest published evidence, and what to validate before choosing it. The marker compares it with the current ranked catalog, not with an absolute quality threshold.
Strongest published evidence
Reasoning ranks #28. Particularly strong for complex reasoning, multi-step problem solving, and analytical tasks.
Validate before choosing
18 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.
Source-linked · 18 displayable benchmark rows
Is Ling 3.1 Flash right for your job?
Its place on the LLM Selector’s job rankings, the same order the job pages publish. Describe your own job to see where it stands for that.
Where Ling 3.1 Flash ranks by benchmark score:
We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank #30 of 135Percentile 78thWeight 22%10 benchmarksVerified | 61.4 | 10 benchmarks | Verified | |||
| CodingRank #29 of 162Percentile 83rdWeight 20%5 benchmarksVerified | 56.8 | 5 benchmarks | Verified | |||
| ReasoningRank #28 of 160Percentile 83rdWeight 17%Not listedUnlisted sources | 61.4 | Not listed | Unlisted sources | |||
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| KnowledgeRank Not rankedWeight 12%2 benchmarksVerified | 56.0 | 2 benchmarks | Verified | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MathWeight 5%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured |
18 of 919 tracked benchmark slots have displayable evidence · bars run 0–100
Coverage detailsHow much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic10/10 verified
- Coding5/5 verified
- ReasoningNot measured
- MultimodalNot measured
- Knowledge2/2 verified
- MultilingualNot measured
- Inst. FollowingNot measured
- MathNot measured
Capability shape
Each axis shows percentile within that category’s eligible cohort. The comparison outline is the median of the six nearest public-score peers; a collapsed vertex means the category is not rank-eligible.
Ling 3.1 Flash category percentile values
- Agentic78th percentile
- Coding83rd percentile
- Reasoning83rd percentile
- MultimodalNot eligible
- KnowledgeNot eligible
- MultilingualNot eligible
- Instruction followingNot eligible
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#30/135
- Coding#29/162
- Reasoning#28/160
- MultimodalNot ranked
- KnowledgeNot ranked
- MultilingualNot ranked
- Inst. FollowingNot ranked
- MathNot ranked
Benchmark ledger
Every result says who ran the test. For this model, InclusionAI ran 17 of 18. Coding opens by default. The tick marks the best source-verified result in the catalog; provisional leaders do not set the reference.
Coding5 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| SWE-bench Pro | 68.7% | Best verified: Claude Fable 5.2 · 93% | Gap24.3 behind | Reported by InclusionAI | Counts26% ref. weight |
| DeepSWE | 59.7% | Best verified: Muse Wonder Full · 79.0% | Gap19.3 behind | Reported by InclusionAI | Counts15% ref. weight |
| SWE Multilingual | 88.5% | Best verified: Claude Fable 5.2 · 96.6% | Gap8.1 behind | Reported by InclusionAI | Counts5% ref. weight |
| terminalBench4 | 46.0% | Best verified: Ling 3.1 Flash · 46.0% | GapHighest listed | Reported by InclusionAI | Shown only |
| SWE-Atlas Codebase QnA | 64.5% | Best verified: Ling 3.1 Flash · 64.5% | GapHighest listed | Reported by InclusionAI | Shown only |
Agentic10 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| Terminal-Bench 2.1 | 81.2% | Best verified: SWE-2 · 92.8% | Gap11.6 behind | Reported by InclusionAI | Counts9% ref. weight |
| BrowseComp | 91.7% | Best verified: Atria Dawn Preview · 92.5% | Gap0.8 behind | Reported by InclusionAI | Counts9% ref. weight |
| AutomationBench | 52.5% | Best verified: MiniMax M3.1 Flash · 62.2% | Gap9.7 behind | Reported by InclusionAI | Counts6% ref. weight |
| MCP Atlas | 83% | Best verified: Muse Spark Open · 89.2% | Gap6.2 behind | Reported by InclusionAI | Counts5% ref. weight |
| τ³-bench resultsτ³-Bench Tool-Agent-User Evaluation | 47.8% | Best verified: Mercury 2.5 · 96.0% | Gap48.2 behind | Reported by InclusionAI | Counts2% ref. weight |
| skillsBench | 68.7% | Best verified: Qwen3.8 Max · 70.2% | Gap1.5 behind | Reported by InclusionAI | Shown only |
| CyberGym | 87.9% | Best verified: MiMo-V2.6-Flash · 95.1% | Gap7.2 behind | Reported by InclusionAI | Shown only |
| Finance Agent v2 | 57.9% | Best verified: Gemini 4 Argon · 65.4% | Gap7.5 behind | Reported by InclusionAI | Shown only |
| DRACOData Research and Analysis with Complex Operations | 85.5% | Best verified: Claude Opus 5 · 88.6% | Gap3.1 behind | Reported by InclusionAI | Shown only |
| WideResearch | 83.3% | Best verified: Hy4 preview · 83.9% | Gap0.6 behind | Reported by InclusionAI | Shown only |
Knowledge2 rows
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| HLEHumanity's Last Exam | 37.4% | Best verified: Claude Fable 5.2 · 74.2% | Gap36.8 behind | Reported by InclusionAI | Counts44% ref. weight |
| HealthBench Professional | 65.3% | Best verified: Claude Sonnet 5.5 · 69.2% | Gap3.9 behind | Reported by InclusionAI | Shown only |
External signals1 row
| Benchmark | Score | Versus best verified row | Gap | Who ran it | Use in ranking |
|---|---|---|---|---|---|
| StealthMarkStealthMark Intelligence Benchmarks (overall score) | 42.4 | Best verified: GPT-6 Astra · 74.6 | Gap32.2 behind | Run by StealthMark | Shown only |
Outlined bars are lab-reported. Filled bars are benchmark-run or independent.
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Flash
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- inclusionai/ling-3.1-flashVercel AI Gateway Ling 3.1 Flash route
- Context window
- 262KInclusionAI Ling 3.1 Flash released configuration
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- textInclusionAI Ling 3.1 Flash model card
- Output modalities
- textInclusionAI Ling 3.1 Flash model card
- Parameters
- Not sourced yet
- Availability
- Hugging Face weights · SGLang · vLLM · Vercel AI Gateway API via NovitaVercel AI Gateway Ling 3.1 Flash route
- Cloud regions
- Not tracked yet
- Lifecycle
- activeInclusionAI Ling 3.1 Flash model card
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Not documented in the pricing recordVercel AI Gateway availability
- Self-host
- Ling 3.1 Flash needs ~640GB VRAM (8× NVIDIA H100 (80GB)).
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
Ling 3.1 Flash ranks #30 of 215 on the public leaderboard with a score of 64.65/100. It does not yet have enough sourced coverage for a verified position.
Ling 3.1 Flash is a open weight model with a 262K context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
Public FP8 weights are available under MIT as inclusionAI/Ling-3.1-flash, with SGLang and vLLM serving instructions. The released configuration sets max_position_embeddings to 262,144 tokens; the model card separately claims up to 1M context. The existing Vercel AI Gateway route via Novita supports 262,144 context tokens and 32,768 output tokens. Its free promotion ended October 13, 2026.
Ling 3.1 Flash has cybersecurity benchmark results. The cybersecurity benchmark comparison shows them beside other models, one security job at a time, with the access route for restricted models.
InclusionAI reports 560B total and 25B active parameters in a hybrid-linear MoE with 72 layers, 512 routed experts, and eight active experts per token. Its current model-card chart reports Terminal-Bench 4.0 at 46.03 and SWE-Atlas Codebase QnA at 64.51, replacing the September launch-chart values. Those results, SWE-bench Pro, and DeepSWE use LingCode, a customized Pi harness. Other harness and grading details are not specified. HealthBench Professional uses the AQ environment; this is provider-reported evidence, not independent testing or clinical certification. FrontierSWE is not mapped because the chart does not identify its version. Publisher claims of 1M context remain separate from the released 262,144-token configuration.
Its explicit predecessor is Ling 3.0 Flash. 18 of 919 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Reasoning at #28, while its lowest eligible position is Agentic at #30. particularly strong for complex reasoning, multi-step problem solving, and analytical tasks.
Last updated November 13, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Where to run Ling 3.1 Flash by API
2 hosts publish a per-token price for it. OpenRouter has the lowest output rate, $0.14 per million tokens. Prices read from each host’s own pages on November 11, 2026.
| Host | Input /1M | Output /1M | Cached /1M | Context |
|---|---|---|---|---|
| OpenRouter | $0.045 | $0.14 | $0.009 | 1M |
| Novita AI | $0.30 | $0.90 | $0.06 | 1M |
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Questions
How does Ling 3.1 Flash perform overall in AI benchmarks?
Ling 3.1 Flash ranks #30 out of 215 models on the public BenchAlign leaderboard, with a score of 64.65/100. Its evidence status is Estimated, and this profile shows 18 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is Ling 3.1 Flash good for knowledge and understanding?
Ling 3.1 Flash has source-displayable benchmark coverage for knowledge and understanding, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is Ling 3.1 Flash good for coding and programming?
Ling 3.1 Flash ranks #29 out of 162 eligible models for coding and programming, with a public category score of 56.8/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Ling 3.1 Flash good for agentic tool use and computer tasks?
Ling 3.1 Flash ranks #30 out of 135 eligible models for agentic tool use and computer tasks, with a public category score of 61.4/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is Ling 3.1 Flash open source?
Ling 3.1 Flash is an open-weight model from InclusionAI. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
Does Ling 3.1 Flash have full benchmark coverage on BenchLM?
No. Ling 3.1 Flash currently has 24 source-displayable rows. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of Ling 3.1 Flash?
Ling 3.1 Flash has a documented context window of 262K. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.