BlogLLM Benchmark Scores as Published, Side by Side
DeepSeek-V3-0324: Benchmark Scores as Published
Every result the catalog holds for DeepSeek-V3-0324: 36 tests in 8 areas, from 3 pages. 5 printed by DeepSeek itself, the rest by the company named beside each number. Each number is as printed; nothing is ranked or rescored.
Developer DeepSeekReleased 2025-03-24Input TextOutput TextWeights Open (MIT)Data version 2026103001, as of 2026-10-30
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| SWE-bench Verified | 38.8% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Harness: agentic, single attempt |
| SWE-bench Multilingual | 25.8% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Harness: agentic, single attempt |
| Terminal-Bench (version not stated) | 16.3% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Harness: Terminus |
| SWE-bench Verified (agentless, single patch without test) | 36.6% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | Harness: agentless |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPQA Diamond | 68.4% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15; printed on 2 more pages | - |
| AIME 2025 | 46.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| HMMT February 2025 | 27.5% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| AIME 2024 | 59.4% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15; printed on 1 more page | - |
| MATH-500 | 94% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15; printed on 2 more pages | - |
| CNMO 2024 | 74.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| PolyMath (en) | 59.5% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| RULER at 128K | 91.9% | Mistral AI, Medium is the new largechecked 2026-10-15 | Context: 128000 |
| RULER at 32K | 95.8% | Mistral AI, Medium is the new largechecked 2026-10-15 | Context: 32000 |
General Knowledge and Reasoning
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Humanity's Last Exam (no tools) | 5.2% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| MMLU-Pro | 81.2% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15; printed on 2 more pages | - |
| MMLU-Redux | 90.5% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| SuperGPQA | 53.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| MMLU | 89.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| SimpleQA | 27.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| AutoLogi | 88.9% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| LiveBench (2024-11-25) | 72.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| ZebraLogic | 84% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
Code Generation
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LiveCodeBench v6 | 42.9% | Mistral AI, Medium is the new largechecked 2026-10-15; printed on 1 more page | - |
| LiveCodeBench (version not stated) | 49.2% | DeepSeek, DeepSeek-V3-0324 model cardchecked 2026-10-15 | - |
| Aider polyglot | 55.1% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| MultiPL-E | 84.9% | Mistral AI, Medium is the new largechecked 2026-10-15; printed on 1 more page | - |
| OJBench (language not stated) | 24% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| HumanEval | 93.3% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| τ²-bench Telecom | 32.5% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| τ²-bench Retail | 69.1% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| τ²-bench Airline | 39% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| AceBench | 72.7% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
Instruction Following
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| MultiChallenge | 31.4% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
| IFEval | 89.1% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
| IFEval (prompt-level strict) | 81.1% | Moonshot AI, Kimi K2 release pagechecked 2026-10-15 | - |
Preference and Composite Rankings
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Arena-Hard win rate (GPT-4-1106 judge) | 97.3% | Mistral AI, Medium is the new largechecked 2026-10-15 | - |
More from DeepSeek
- DeepSeek-V4.1-Flash-1017
- DeepSeek-V4.1-Flash
- DeepSeek-V4-Flash (Preview)
- DeepSeek-V4-Flash-Base
- DeepSeek-V4-Pro (Preview)
- DeepSeek-V4-Pro-Base
- DeepSeek-V3.2
- DeepSeek-V3.2-Base
- DeepSeek-V3.2-Speciale
- DeepSeek-V3.2-Exp
- DeepSeek-V3.1-Terminus
- DeepSeek-V3.1
- DeepSeek-R1-0528
- DeepSeek-R1-0528-Qwen3-8B
- DeepSeek-R1
- DeepSeek-R1-Distill-Llama-70B
- DeepSeek-R1-Distill-Llama-8B
- DeepSeek-R1-Distill-Qwen-1.5B
- DeepSeek-R1-Distill-Qwen-14B
- DeepSeek-R1-Distill-Qwen-32B
- DeepSeek-R1-Distill-Qwen-7B
- DeepSeek-V3
- DeepSeek-V4-Flash-0731
- DeepSeek-V4-Pro-0813
- DeepSeek-V4-Vision-Exp
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.