DeepSeekDeepSeek V4.1 Flash
DeepSeek V4.1 Flash is a DeepSeek open-weights reasoning model, released 10 Sept 2026. Its best configuration (max reasoning effort) ranks #89 of 756 on the BenchLeader Index at 61.7 ±2.8, in the upper half. The ± is the point: 160 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in composite (71) and lowest in agents & tools (56). At $0.262 per million tokens blended it is cheaper than most ranked models. Output speed of 209 tokens per second puts it in the fastest quarter, with a first answer in 10.5 s. It has been measured at 4 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 2 Oct 2026.
- Blended price
- $0.262/M
- $0.150 in · $0.600 out
- Output speed
- 209 tok/s
- measured by Artificial Analysis
- First answer
- 11 s
- first token 1.01 s
- Context
- 1M
- Full answer
- 13 s
- median, reasoning included
- Cost per run
- $0.265
- one full Intelligence Index run
- Released
- 10 Sept 2026
- Overall index62
- Reasoning66
- Coding64
- Agents & tools56
- Maths69
- Knowledge59
- Human preference67
- Multimodal60
- Long context68
- Composite71
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | 49.2 | #390 | 215 tok/s | 0.96 s | $0.0002 | Coding 35 · Composite 58 · Knowledge 57 · Long context 53 · Multimodal 48 · Reasoning 42 |
| high | – | – | 97 tok/s | 1.01 s | $0.0002 | Agents & tools 61 · Coding 39 · Knowledge 58 |
| maxbest | 61.7 | #89 | 209 tok/s | 11 s | $0.0002 | Agents & tools 56 · Coding 64 · Composite 71 · Human preference 67 · Knowledge 59 · Long context 68 · Maths 69 · Multimodal 60 · Reasoning 66 |
| not stated | – | – | 97 tok/s | 1.01 s | $0.0002 | Agents & tools 39 · Maths 57 |
How its index has moved
19 Sept 2026 to 5 Oct 2026- max
- no reasoning
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Metadata
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| AA Openness Indexnot in index | – | – | 44.4#26 | – | Artificial Analysis |
Reasoning
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| LMArena Hard Prompts | – | – | 1500#34 | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | 86.7%#30 | – | LiveBench | |
| Humanity's Last Exam (AA)not in index | 10.8%#290 | – | 39.3%#84 | – | Artificial Analysis | |
| CritPt | 0.3%#250 | – | 14.3%#87 | – | Artificial Analysis | |
| MysteryMechanismnot in index | – | 21.2%#17 | – | – | Vals AI | |
| LMCAnot in index | – | – | – | 47.0%#66 | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | – | – | – | 89.9%#73 | Epoch AI Benchmarking Hub |
Coding
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| SciCode | 35.5%#180 | – | 51.9%#80 | – | SciCode | |
| LMArena Coding | – | – | 1528#26 | – | LMArena | |
| LMArena WebDev | – | – | 1620#20 | – | LMArena | |
| LiveBench Codingnot in index | – | – | 80.0%#21 | – | LiveBench | |
| SciCode (AA)not in index | 35.5%#182 | – | 51.9%#81 | – | Artificial Analysis | |
| IOI | – | 40.3%#29 | – | – | Vals AI | |
| Code Migrationnot in index | – | 45.6%#14 | – | – | Vals AI | |
| Vibe Code Bench 1-100not in index | – | 16.4%#12 | – | – | Vals AI | |
| Vibe Code Bench v1.1not in index | – | 84.7%#14 | – | – | Vals AI | |
| Surface Evolver Benchnot in index | – | 46.3%#20 | – | – | Epoch AI Benchmarking Hub | |
| ALE-Benchnot in index | – | – | 1092.3#45 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | 5.6%#86 | – | 26.8%#42 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| APEX-Agents | – | – | – | 39.5%#40 | Mercor | |
| LMArena Agent | – | – | 4#16 | – | LMArena | |
| LiveBench Agentic Codingnot in index | – | – | 77.3%#1 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | – | 74.5%#18 | – | – | Vals AI | |
| GDPval (AA)not in index | 41.4%#83 | – | 55.0%#24 | – | Artificial Analysis | |
| ITBench SRE (AA)not in index | – | – | 46.9%#10 | – | Artificial Analysis | |
| CyberBenchnot in index | – | 73.7%#8 | – | – | Vals AI | |
| Finance Agent v2not in index | – | 53.5%#28 | – | – | Vals AI | |
| Harvey's Legal Agent Benchmarknot in index | – | – | 6.7%#23 | – | Vals AI | |
| Legal Research Benchnot in index | – | 41.4%#20 | – | – | Vals AI | |
| SkillsBenchnot in index | – | 69.8%#1 | – | – | Vals AI | |
| SREBenchnot in index | – | – | 0.8%#13 | – | Vals AI | |
| Tax Agent Benchnot in index | – | 62.5%#25 | – | – | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in index | – | 19.7%#19 | – | – | Vals AI | |
| Terminal-Bench Sciencenot in index | – | 0.0%#29 | – | – | Vals AI | |
| AutomationBenchnot in index | – | – | 68.9%#4 | – | Artificial Analysis | |
| GDP.pdfnot in index | – | – | 12.8%#37 | – | Artificial Analysis | |
| AA-Briefcasenot in index | – | – | 1420#34 | – | Artificial Analysis |
Maths
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| ProofBench | – | – | – | 54.0%#25 | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | 93.3%#21 | – | LiveBench | |
| LMArena Maths | – | – | 1503#14 | – | LMArena |
Knowledge
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| LiveBench Data Analysisnot in index | – | – | 79.3%#16 | – | LiveBench | |
| AA-Omniscience | -10.2#177 | – | -5.3#151 | – | Artificial Analysis | |
| LegalBench | – | 83.3%#53 | – | – | Vals AI | |
| BioMysteryBenchnot in index | – | 67.8%#13 | – | – | Vals AI | |
| Excel Modeling Benchmarknot in index | – | 57.2%#35 | – | – | Vals AI | |
| MedCodenot in index | – | 41.2%#53 | – | – | Vals AI | |
| MedScribenot in index | – | 85.5%#25 | – | – | Vals AI | |
| Public Benefits Benchnot in index | – | – | 64.3%#19 | – | Vals AI | |
| GDP.pdfnot in index | – | – | 19.8%#27 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 28.4%#180 | – | 46.4%#79 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 46.0%#125 | – | 3.5%#517 | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | 81.2%#27 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | 70.0%#29 | – | LiveBench | |
| LMArena Instruction Followingnot in index | – | – | 1479#21 | – | LMArena |
Human preference
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| LMArena Text | – | – | 1474#38 | – | LMArena | |
| LMArena Creative Writingnot in index | – | – | 1440#59 | – | LMArena | |
| LMArena Multi-turnnot in index | – | – | 1461#73 | – | LMArena | |
| LMArena Longer Queriesnot in index | – | – | 1482#41 | – | LMArena |
Multimodal
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| MMMU-Pro | 65.6%#168 | – | 77.0%#81 | – | Artificial Analysis | |
| SAGEnot in index | – | 47.9%#29 | – | – | Vals AI |
Long context
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| AA-LCR | 55.3%#277 | – | 84.0%#11 | – | Artificial Analysis | |
| MLCRnot in index | – | – | 22.8%#15 | – | Artificial Analysis |
Composite
| Benchmark | no reasoning | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | 155#28 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | 81.1%#8 | – | LiveBench | |
| AA Intelligence Index | 24.7#157 | – | 39.5#61 | – | Artificial Analysis | |
| Vals Indexnot in index | – | 51.3#23 | – | – | Vals AI |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Together | 219 tok/s | 0.29 s | $0.300 | $1.20 | 1.0M | – |
| Baseten | 203 tok/s | 0.40 s | $0.300 | $1.20 | 1.0M | fp8 |
| Baseten | 189 tok/s | 0.40 s | $0.300 | $1.20 | 1.0M | fp8 |
| Modal | 185 tok/s | 0.89 s | $0.300 | $1.20 | 1.0M | – |
| Makora | 173 tok/s | 0.80 s | $0.270 | $1.15 | 1.0M | fp8 |
| Fireworks | 150 tok/s | 1.36 s | $0.450 | $1.80 | 1.0M | – |
| DeepSeek | 132 tok/s | 1.19 s | $0.150 | $0.600 | 1.0M | – |
| CoreWeave | 122 tok/s | 0.65 s | $0.200 | $0.650 | 1.0M | fp8 |
| Baidu Qianfan | 112 tok/s | 1.04 s | $0.300 | $1.20 | 1.0M | fp8 |
| NovitaAI | 99 tok/s | 2.03 s | $0.240 | $0.960 | 1.0M | fp8 |
| Relace | 94 tok/s | 0.92 s | $0.003 | $2.40 | 1.0M | – |
| Venice | 94 tok/s | 1.41 s | $0.300 | $1.20 | 1M | fp8 |
| AtlasCloud | 92 tok/s | 1.22 s | $0.141 | $0.564 | 1.0M | fp8 |
| GMICloud | 88 tok/s | 2.36 s | $0.180 | $0.720 | 1.0M | fp8 |
| NextBit | 88 tok/s | 2.78 s | $0.210 | $0.840 | 1.0M | fp8 |
| SiliconFlow | 85 tok/s | 2.63 s | $0.300 | $1.20 | 1.0M | fp8 |
| StreamLake | 76 tok/s | 1.82 s | $0.141 | $0.564 | 1.0M | fp8 |
| Ionstream | 73 tok/s | 1.33 s | $0.119 | $1.18 | 1.0M | – |
| Alibaba Cloud Int. | 71 tok/s | 1.25 s | $0.150 | $0.600 | 1M | – |
| Decart | 69 tok/s | 1.23 s | $0.090 | $0.180 | 1.0M | fp4 |
| Wafer | 68 tok/s | 0.78 s | $0.075 | $1.20 | 1.0M | – |
| DeepInfra | 66 tok/s | 1.03 s | $0.140 | $0.420 | 1.0M | fp8 |
| Phala | 55 tok/s | 2.04 s | $0.210 | $0.840 | 1.0M | – |
| inference.net | 53 tok/s | 1.11 s | $0.070 | $0.300 | 1.0M | – |
| Fireworks | 53 tok/s | 3.03 s | $0.300 | $1.20 | 1.0M | – |
| DigitalOcean | 48 tok/s | 0.92 s | $0.180 | $0.720 | 1.0M | – |
| DekaLLM | 47 tok/s | 1.43 s | $0.120 | $2.40 | 1.0M | – |
| Sail Research | 28 tok/s | 2.00 s | $0.080 | $0.400 | 1.0M | fp4 |
| Morph | 20 tok/s | 2.44 s | $0.058 | $0.450 | 1.0M | fp8 |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.003 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0002 | $0.0002 | 11.9 s |
| Summarise a 30-page report | 12,000 / 600 | $0.0022 | $0.0008 | 13.4 s |
| Code edit | 6,000 / 1,500 | $0.0018 | $0.0011 | 17.7 s |
| Agentic coding session | 60,000 / 4,000 | $0.011 | $0.0048 | 29.6 s |
| Structured extraction | 2,000 / 200 | $0.0004 | $0.0002 | 11.5 s |
See also
Data as of 5 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “DeepSeek V4.1 Flash: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/deepseek-v4-1-flash, data as of 5 Oct 2026.