DeepSeek V4.1 Flash
DeepSeek V4.1 Flash model for reasoning and agentic coding.
- Capability
- 157.0ECI · #34 of 158
- Input
- $0.232per 1M tokens
- Output
- $1.08per 1M tokens
- Context
- 1M384K max output
What it costs
- Input
- $0.232per million tokens
- Output
- $1.08per million tokens
- Blended (3:1)
- $0.444cheaper than 63% of priced models
Typical monthly bills
| Monthly usage | Estimated cost |
|---|---|
| Side project2M input + 0.5M output tokens | $1.00 |
| Team assistant25M input + 5M output tokens | $11.19 |
| Production app250M input + 50M output tokens | $111.88 |
Median of 82 paid API providers on models.dev.
Independent benchmarks
#34 of 158 scored models
The shaded band is Epoch AI’s confidence range (155.1–159.0); the tick marks the median scored model.
Scores from Epoch AI, run independently of DeepSeek.
The details
- Lab
- DeepSeek
- Released
- Sep 10, 2026
- Knowledge cutoff
- May 2025
- Context window
- 1,000,000 tokens
- Max output
- 384,000 tokens
- Inputs
- Text, Images
- Output
- Text
- Reasoning
- Always on
- Tool calling
- Yes
- Structured output
- No
- Weights
- Open weights · MIT
- Availability
- 86 API providerslisted on models.dev
What DeepSeek claims
Published by the lab at launch. Settings vary, so compare these only with care.
| Benchmark | Score | Setting | Source |
|---|---|---|---|
| GPQA Diamond | 90.9 | reasoning_effort=100 | Source |
| MathArena Apex | 65.6 | reasoning_effort=100 | Source |
| Humanity's Last Exam | 36.8 | reasoning_effort=100 | Source |
| Humanity's Last Exam | 39.1 | reasoning_effort=100 | Source |
| Codeforces | 3471 rating | reasoning_effort=100 | Source |
| Terminal-Bench v2.1 | 90.6 | reasoning_effort=100 | Source |
| Terminal-Bench v3.0 | 30 | reasoning_effort=100 | Source |
| Terminal-Bench v4.0 | 31.2 | reasoning_effort=100 | Source |
| DeepSWE v1.1 | 74.2 resolved | reasoning_effort=100 | Source |
| ProgramBench | 20.3 almost@1 | reasoning_effort=100 | Source |
| NL2Repo | 64 | reasoning_effort=100 | Source |
| CyberGym | 88.1 | reasoning_effort=100 | Source |
More from DeepSeek
About DeepSeek V4.1 Flash
How much does DeepSeek V4.1 Flash cost?
DeepSeek V4.1 Flash costs $0.232 input / $1.08 output per million tokens (median across 82 API providers). At a 3:1 input-to-output mix that is $0.444 per million tokens, cheaper than 63% of the 469 priced models we track.
What is the context window of DeepSeek V4.1 Flash?
DeepSeek V4.1 Flash accepts up to 1,000,000 tokens per request and can write up to 384,000 tokens in one response.
How good is DeepSeek V4.1 Flash?
Epoch AI gives DeepSeek V4.1 Flash a Capabilities Index score of 157.0 (likely range 155.1–159.0), ranking it #34 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 89.8%, FrontierMath Tiers 1–3 67.4%, OTIS Mock AIME 2024–2025 98.3%, SimpleQA Verified 42.9%.
Is DeepSeek V4.1 Flash open source?
Yes. DeepSeek publishes the weights under the MIT licence, so it can be downloaded and self-hosted.
What inputs does DeepSeek V4.1 Flash support?
DeepSeek V4.1 Flash accepts text and images and replies in text. It is a reasoning model, supports tool calling.
When was DeepSeek V4.1 Flash released?
DeepSeek released DeepSeek V4.1 Flash on Sep 10, 2026. Its training data runs to May 2025.
What are the best alternatives to DeepSeek V4.1 Flash?
The closest current models from other labs on capability, price and release date are GPT-5.6 Luna (OpenAI, ECI 156.7, $0.20 / $1.20), Seed 2.1 Turbo (ByteDance Seed, $0.445 / $2.23), GLM-5.3-Flash (Z.ai (Zhipu), ECI 153.8, $0.15 / $0.50) and Mistral Large 4 (Mistral AI, ECI 154.6, $0.70 / $2.10).