Skip to main content
benchlm.ai

DeepSeek V3.1 (Reasoning)

DeepSeekEstablishedReleased Aug 21, 2025
Overall Score
Est. 29Prov. #94 of 119
Arena Elo
1418
Categories Ranked
8of 8
Price (1M tokens)
$0 in / $0 out
Speed
N/A
Context
128K
Open WeightSelf-hostReasoning
Confidence
reasoning

According to BenchLM.ai, DeepSeek V3.1 (Reasoning) ranks #94 out of 119 models on the provisional leaderboard with an overall score of 29/100. It does not yet have enough sourced coverage for BenchLM's verified leaderboard. While not a frontier model, it offers specific advantages depending on the use case.

DeepSeek V3.1 (Reasoning) is a open weight model with a 128K token context window. It uses explicit chain-of-thought reasoning, which typically improves performance on math and complex reasoning tasks at the cost of higher latency and token usage.

DeepSeek V3.1 (Reasoning) sits inside the DeepSeek V3.1 family alongside DeepSeek V3.1. This profile currently has 16 of 225 tracked benchmarks. BenchLM only exposes non-generated benchmark rows publicly, so missing categories stay blank until a sourced evaluation is available.

Its strongest category is Agentic (#68), while its weakest is Instruction Following (#110). This performance profile makes it particularly useful for coding agents, browser research, and computer-use workflows.

Ranking Distribution

Category rank across 8 benchmark categories — sorted by best rank

Category Performance

Scores across all benchmark categories (0-100 scale)

Category Breakdown

Agentic

#68
33.6/ 100
Weight: 22%4 benchmarks
Terminal-Bench 2.0BrowseCompOSWorld-VerifiedGAIATAU-benchWebArena

Coding

#77
21.7/ 100
Weight: 20%3 benchmarks
SWE-bench VerifiedLiveCodeBenchSWE-bench ProSWE-RebenchSciCode

Reasoning

#76
26.8/ 100
Weight: 17%2 benchmarks
MuSRLongBench v2MRCRv2ARC-AGI-2

Knowledge

#96
13.8/ 100
Weight: 12%6 benchmarks
GPQASuperGPQAMMLU-ProHLEFrontierScienceSimpleQA

Math

#82
7.5/ 100
Weight: 5%0 benchmarks
AIME 2025BRUMO 2025MATH-500FrontierMath

Multilingual

#83
11.5/ 100
Weight: 7%0 benchmarks
MGSMMMLU-ProX

Multimodal

#99
6.2/ 100
Weight: 12%0 benchmarks
MMMU-ProOfficeQA ProCharXivCharXiv w/o tools

Inst. Following

#110
14.1/ 100
Weight: 5%1 benchmark
IFEvalIFBench

Chatbot Arena Performance

Text Overall1418CI: ±6.015,080 votes
Coding1447CI: ±11.62,635 votes
Math1416CI: ±18.2998 votes
Instruction Following1403CI: ±9.83,718 votes
Creative Writing1390CI: ±13.22,028 votes
Multi-turn1406CI: ±11.82,486 votes
Hard Prompts1433CI: ±7.86,837 votes
Hard Prompts (English)1436CI: ±10.23,381 votes
Longer Query1421CI: ±10.53,247 votes

Benchmark Details

Only benchmark rows with an attached exact-source record are shown here. Source-unverified manual rows and generated rows are hidden from model pages.

DeepSeek V3.1 Family

Reasoning

Canonical Entry

DeepSeek V3.1
DeepSeek V3.1Prov. 25

Compare This Model

See how DeepSeek V3.1 (Reasoning) stacks up against similar models

DeepSeek V3.1DeepSeek
Prov. 25
DeepSeek V4 Pro (Max)DeepSeek
Prov. 87
DeepSeek V4 Pro (High)DeepSeek
Prov. 83
DeepSeek V4 Flash (Max)DeepSeek
Prov. 75
Claude Mythos PreviewAnthropic
Prov. 99

Frequently Asked Questions

How does DeepSeek V3.1 (Reasoning) perform overall in AI benchmarks?

DeepSeek V3.1 (Reasoning) currently ranks #94 out of 119 models on BenchLM's provisional leaderboard with an overall score of 29 (estimated). It is created by DeepSeek and features a 128K context window.

Is DeepSeek V3.1 (Reasoning) good for knowledge and understanding?

DeepSeek V3.1 (Reasoning) ranks #96 out of 119 models in knowledge and understanding benchmarks with an average score of 13.8. There are stronger options in this category.

Is DeepSeek V3.1 (Reasoning) good for coding and programming?

DeepSeek V3.1 (Reasoning) ranks #77 out of 119 models in coding and programming benchmarks with an average score of 21.7. There are stronger options in this category.

Is DeepSeek V3.1 (Reasoning) good for reasoning and logic?

DeepSeek V3.1 (Reasoning) ranks #76 out of 119 models in reasoning and logic benchmarks with an average score of 26.8. There are stronger options in this category.

Is DeepSeek V3.1 (Reasoning) good for agentic tool use and computer tasks?

DeepSeek V3.1 (Reasoning) ranks #68 out of 119 models in agentic tool use and computer tasks benchmarks with an average score of 33.6. There are stronger options in this category.

Is DeepSeek V3.1 (Reasoning) good for instruction following?

DeepSeek V3.1 (Reasoning) ranks #110 out of 119 models in instruction following benchmarks with an average score of 14.1. There are stronger options in this category.

Is DeepSeek V3.1 (Reasoning) open source?

Yes, DeepSeek V3.1 (Reasoning) is an open weight model created by DeepSeek, meaning it can be downloaded and run locally or fine-tuned for specific use cases.

Which sibling models are related to DeepSeek V3.1 (Reasoning)?

DeepSeek V3.1 (Reasoning) belongs to the DeepSeek V3.1 family. Related variants on BenchLM include DeepSeek V3.1.

Does DeepSeek V3.1 (Reasoning) have full benchmark coverage on BenchLM?

Not yet. DeepSeek V3.1 (Reasoning) currently has 16 published benchmark scores out of the 225 benchmarks BenchLM tracks. BenchLM only exposes non-generated public benchmark rows, so missing categories stay blank until a sourced evaluation is available.

What is the context window size of DeepSeek V3.1 (Reasoning)?

DeepSeek V3.1 (Reasoning) has a context window of 128K, which determines how much text it can process in a single interaction.

Related Resources

More from DeepSeek LLM Pricing Comparison BenchLM Methodology The AI Race Which LLM Should I Use? Benchmark Explainers
Last updated: June 2, 2026 · Runtime metrics stay blank until BenchLM has a sourced snapshot.

Don't miss the next GPT moment

Which models moved up, what’s new, and what it costs. One email a week, 3-min read.

Free. One email per week.