Data as of October 20, 2026 · How the score is built
Released Jul 23, 2026262K contextInclusionAI Ling 3.0 Flash model card
Ling 3.0 Flash
Decision readingLing 3.0 Flash scores 45.1 out of 100 and ranks #125 of 227. This profile shows 24 source-displayable benchmark rows; its strongest eligible category is Instruction Following at #68. No comparable first-party API price is published in the catalog.
Released Jul 23, 2026 — see all recent releases
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Capability
45.1/100
field median 49.8#125 of 227 ranked models
Public
#125of 227
Verified —
Price
API rate not published
input median $0.72No comparable first-party hosted token rate
Speed
285tok/s
field median 93.5 tok/sFirst token 9.81 s
Context
262Ktokens
field median 256,000Maximum output length is tracked separately
Strongest published evidence
Instruction Following ranks #68. A well-rounded choice across a range of tasks.
Validate before choosing
24 published rows leave some tracked benchmark slots empty. No comparable first-party API token rate is published.
Source-linked · 24 displayable benchmark rows
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 #87 of 131Percentile 34thWeight 22%9 benchmarksVerified | 33.4 | 9 benchmarks | Verified | |||
| CodingRank #92 of 152Percentile 40thWeight 20%6 benchmarksVerified | 31.6 | 6 benchmarks | Verified | |||
| ReasoningWeight 17%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| KnowledgeRank #107 of 187Percentile 43rdWeight 12%5 benchmarksVerified | 43.5 | 5 benchmarks | Verified | |||
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | 0 benchmarks | Not measured | |||
| Inst. FollowingRank #68 of 149Percentile 55thWeight 5%1 benchmarkVerified | 75.2 | 1 benchmark | Verified | |||
| MathRank Not rankedWeight 5%3 benchmarksVerified | 81.1 | 3 benchmarks | Verified |
24 of 726 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.
- Agentic9/9 verified
- Coding6/6 verified
- ReasoningNot measured
- MultimodalNot measured
- Knowledge5/5 verified
- MultilingualNot measured
- Inst. Following1/1 verified
- Math3/3 verified
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 3.0 Flash category percentile values
- Agentic34th percentile
- Coding40th percentile
- ReasoningNot eligible
- MultimodalNot eligible
- Knowledge43rd percentile
- MultilingualNot eligible
- Instruction following55th percentile
- MathNot eligible
The dashed outline is median of 6 nearest peers.
Eligible category ranks
- Agentic#87/131
- Coding#92/152
- ReasoningNot ranked
- MultimodalNot ranked
- Knowledge#107/187
- MultilingualNot ranked
- Inst. Following#68/149
- MathNot ranked
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding6 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| SWE-bench Pro | Score56.6% | Versus best verified row Best verified: Claude Opus 5.5 · 89.9% | Gap33.3 behind | Weight26% ref. weight | Provider exact |
| SciCodeScientific Code Benchmark | Score41.2% | Versus best verified row Best verified: Sakana Fugu · 60.1% | Gap18.9 behind | Weight10% ref. weight | Provider exact |
| LiveCodeBench (Vals)LiveCodeBench, Vals AI run | Score84.0% | Versus best verified row Best verified: Claude Fable 5.1 · 90.5% | Gap6.5 behind | Weight8% ref. weight | |
| SWE Multilingual | Score72.4% | Versus best verified row Best verified: Claude Opus 5.5 · 93.9% | Gap21.5 behind | Weight5% ref. weight | Provider exact |
| LiveCodeBench v5 | Score82.8% | Versus best verified row Best verified: Ling 3.0 Flash · 82.8% | GapBest verified | WeightDisplay only | Provider exact |
| SWE-bench (Vals)SWE-bench, Vals AI run | Score65.2% | Versus best verified row Best verified: Claude Opus 5 · 97.0% | Gap31.8 behind | WeightDisplay only | Verified |
Agentic9 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| BrowseComp | Score72.2% | Versus best verified row Best verified: Atria Dawn Preview · 92.5% | Gap20.3 behind | Weight8% ref. weight | Provider exact |
| MCP Atlas | Score65.5% | Versus best verified row Best verified: Muse Spark Open · 89.2% | Gap23.7 behind | Weight4% ref. weight | Provider exact |
| Terminal-Bench 2.1 (Vals)Terminal-Bench 2.1, Vals AI run | Score50.2% | Versus best verified row Best verified: GPT-6 Astra · 87.3% | Gap37.1 behind | Weight4% ref. weight | |
| BFCL v4Berkeley Function Calling Leaderboard v4 | Score73.0% | Versus best verified row Best verified: BTL-3 · 88.5% | Gap15.5 behind | Weight3% ref. weight | Provider exact |
| skillsBench | Score44.8% | Versus best verified row Best verified: Qwen3.8 Max · 70.2% | Gap25.4 behind | WeightDisplay only | Provider exact |
| WideResearch | Score73.6% | Versus best verified row Best verified: Hy4 preview · 83.9% | Gap10.3 behind | WeightDisplay only | Provider exact |
| DRACOData Research and Analysis with Complex Operations | Score70.4% | Versus best verified row Best verified: Claude Opus 5 · 88.6% | Gap18.2 behind | WeightDisplay only | Provider exact |
| Hermes IndexHermes Index: agentic capability and cost per task inside Hermes Agent | Score21.6% | Versus best verified row Best verified: Claude Opus 5.5 · 63.3% | Gap41.8 behind | WeightDisplay only | Benchmark exact |
| Hermes BenchHermes Bench: 150 agent tasks graded on the files and state the agent leaves behind | Score53.8% | Versus best verified row Best verified: Claude Opus 5.5 · 76.7% | Gap22.9 behind | WeightDisplay only | Benchmark exact |
Knowledge5 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| HLEHumanity's Last Exam | Score22.7% | Versus best verified row Best verified: Claude Fable 5.1 · 65% | Gap42.3 behind | Weight44% ref. weight | Provider exact |
| MMLU-Pro (Vals)MMLU-Pro, Vals AI run | Score82.0% | Versus best verified row Best verified: Claude Fable 5.1 · 92.4% | Gap10.4 behind | Weight6% ref. weight | Verified |
| GPQAGraduate-Level Google-Proof Q&A | Score85.0% | Versus best verified row Best verified: GPT-6 Astra · 96% | Gap11 behind | Weight3% ref. weight | Provider exact |
| GPQA Diamond (Vals)GPQA Diamond, Vals AI run | Score84.8% | Versus best verified row Best verified: Gemini 3.1 Pro · 95.5% | Gap10.7 behind | Weight2% ref. weight | |
| GPQA-DGPQA Diamond | Score85.0% | Versus best verified row Best verified: GPT-6 Astra · 96.0% | Gap11 behind | WeightDisplay only | Provider exact |
Inst. Following1 row
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| IFBenchInstruction Following Benchmark | Score74.5% | Versus best verified row Best verified: MAI-Thinking-1 · 85% | Gap10.5 behind | WeightWeighted 70% | Provider exact |
Math3 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| AIME26AIME 2026 | Score93.2% | Versus best verified row Best verified: GLM-5.2 · 99.2% | Gap6 behind | WeightWeighted 25% | Provider exact |
| HMMT Feb 2026Harvard-MIT Mathematics Tournament February 2026 | Score87.0% | Versus best verified row Best verified: Qwen3.7 Max · 97.1% | Gap10.1 behind | WeightWeighted 25% | Provider exact |
| IMOAnswerBench | Score83.7% | Versus best verified row Best verified: dots3-note Preview · 90.9% | Gap7.2 behind | WeightDisplay only | Provider exact |
Bars run 0–100; the dark tick marks the best source-verified value
All 24 rowsLineage
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.
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
- Not publishedInclusionAI Ling 3.0 Flash model card
- Context window
- 262KInclusionAI Ling 3.0 Flash model card
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- Not sourced yet
- Output modalities
- Not sourced yet
- Parameters
- Not sourced yet
- Availability
- InclusionAI publishes the BF16 checkpoint under the MIT License on Hugging Face. The model card documents SGLang and vLLM deployment, thinking mode enabled by default, and a 262,144-token serving context.
- Cloud regions
- Not tracked yet
- Lifecycle
- Superseded
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Not documented in the pricing recordInclusionAI Ling 3.0 Flash model card
- Self-host
- Open weights available; hardware estimate not sourced
- 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 3.0 Flash ranks #125 of 227 on the public leaderboard with a score of 45.07/100. It does not yet have enough sourced coverage for a verified position.
Ling 3.0 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.
InclusionAI publishes the BF16 checkpoint under the MIT License on Hugging Face. The model card documents SGLang and vLLM deployment, thinking mode enabled by default, and a 262,144-token serving context.
Official exact-value snapshot from InclusionAI's Ling-3.0-flash Hugging Face model card. The final checkpoint is a 124B-total, 5.1B-active hybrid-linear MoE with a 256K training schedule and 262,144-token serving context. BenchLM keeps Terminal-Bench 2.1 separate from the weighted Terminal-Bench 2.0 lane, stores the card's July 2024-May 2025 LiveCodeBench window on the v5 lane, and maps only named rows supported by the current schema. BrowseComp uses the single-agent with-context result; the separate multi-agent result is not collapsed into it. Internal or currently unsupported rows remain outside the scored fields.
Ling 3.0 Flash sits in the Ling 3.0 family with Ling 3.0 Flash VL, Ling 3.0 Flash FP8, Ling 3.0 Tiny, Ling 3.0 Flash Fin. Its explicit predecessor is Ling 2.6 Flash. 24 of 726 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Instruction Following at #68, while its lowest eligible position is Knowledge at #107. a well-rounded choice across a range of tasks.
Last updated October 20, 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 3.0 Flash by API
3 hosts publish a per-token price for it. OpenRouter has the lowest output rate, $0.063 per million tokens. Prices read from each host’s own pages on October 20, 2026.
| Host | Input /1M | Output /1M | Cached /1M | Context |
|---|---|---|---|---|
| OpenRouter | $0.021 | $0.063 | $0.004 | 262K |
| DeepInfra | $0.06 | $0.18 | $0.012 | 131K |
| Novita AI | $0.06 | $0.18 | $0.012 | 262K |
Published weights are available, but BenchLM does not yet have a sourced parameter and VRAM profile for this exact model. Hardware cost estimates stay unavailable until that sizing record is complete.
Estimate VRAM from known parametersQuestions
How does Ling 3.0 Flash perform overall in AI benchmarks?
Ling 3.0 Flash ranks #125 out of 227 models on the public BenchAlign leaderboard, with a score of 45.07/100. Its evidence status is Estimated, and this profile shows 24 source-displayable benchmark rows. The label describes evidence depth, not a provider quality claim; inspect category rows before choosing a workload.
Is Ling 3.0 Flash good for knowledge and understanding?
Ling 3.0 Flash ranks #107 out of 187 eligible models for knowledge and understanding, with a public category score of 43.5/100. 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 3.0 Flash good for coding and programming?
Ling 3.0 Flash ranks #92 out of 152 eligible models for coding and programming, with a public category score of 31.6/100. 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 3.0 Flash good for mathematics?
Ling 3.0 Flash has source-displayable benchmark coverage for mathematics, 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 3.0 Flash good for agentic tool use and computer tasks?
Ling 3.0 Flash ranks #87 out of 131 eligible models for agentic tool use and computer tasks, with a public category score of 33.4/100. 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 3.0 Flash good for instruction following?
Ling 3.0 Flash ranks #68 out of 149 eligible models for instruction following, with a public category score of 75.2/100. 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 3.0 Flash open source?
Ling 3.0 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.
Which sibling models are related to Ling 3.0 Flash?
Ling 3.0 Flash belongs to the Ling 3.0 family. Related tracked variants include Ling 3.0 Flash VL, Ling 3.0 Flash FP8, Ling 3.0 Tiny, plus 1 more. A sibling link indicates shared lineage or a documented configuration relationship; it does not mean the variants have identical pricing, context limits, benchmark evidence, or deployment behavior. Compare before switching.
Does Ling 3.0 Flash have full benchmark coverage on BenchLM?
No. Ling 3.0 Flash currently has 40 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 3.0 Flash?
Ling 3.0 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.