Search
Search titles only
By:
Search titles only
By:
Log in
Register
Search
Search titles only
By:
Search titles only
By:
Menu
Install the app
Install
Forums
New posts
All threads
Latest threads
New posts
Trending threads
Trending
Search forums
What's new
New posts
New ads
New profile posts
Latest activity
Free Ads
Latest reviews
Search ads
Members
Current visitors
New profile posts
Search profile posts
Contact us
Latest ads
Google Pixel 9 Pro
vgp
Updated:
Thursday at 5:57 PM
Ad icon
Jobreceive.com for sale
Blogerwiki
Updated:
Tuesday at 8:55 PM
Post Your Vehicle for Sale — FREE! - https://libro.lk
Kalu_Puth
Updated:
Tuesday at 7:38 PM
ඔයාගෙ Assignment හෝ Thesis එක හරියට හදාගමු
ErMurazor
Updated:
Sep 26, 2026
Ad icon
BlackWall V2ray servers
hu KANNA
Updated:
Sep 23, 2026
Electronics
Vehicles
Property
Search
Reply to thread
Forums
General
ElaKiri Talk!
Introducing DeepSeek-V3: Your AI Companion
Get the App
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser
.
Message
<blockquote data-quote="MihiCherub" data-source="post: 30496402" data-attributes="member: 238676"><p>The DeepSeek R1 model is a versatile, powerful AI tool that’s set to revolutionize how businesses and individuals interact with technology. Whether you’re automating workflows or enhancing user experiences, the R1 is a game-changer.</p><p></p><p><strong>Open-Source Model You Can Run Locally</strong></p><p></p><p><img src="https://dp-cdn-deepseek.obs.cn-east-3.myhuaweicloud.com/api-docs/deepseek_r1_example_en.gif" alt="" class="fr-fic fr-dii fr-draggable " style="width: 771px" /></p><p></p><p><img src="https://huggingface.co/deepseek-ai/DeepSeek-R1/resolve/main/figures/benchmark.jpg" alt="" class="fr-fic fr-dii fr-draggable " style="width: 770px" /></p><h3></h3><h3>DeepSeek-R1-Evaluation </h3> <table style='width: 100%'><tr><th>Category</th><th>Benchmark (Metric)</th><th>Claude-3.5-Sonnet-1022</th><th>GPT-4o 0513</th><th>DeepSeek V3</th><th>OpenAI o1-mini</th><th>OpenAI o1-1217</th><th>DeepSeek R1</th></tr><tr><td></td><td>Architecture</td><td>-</td><td>-</td><td>MoE</td><td>-</td><td>-</td><td>MoE</td></tr><tr><td></td><td># Activated Params</td><td>-</td><td>-</td><td>37B</td><td>-</td><td>-</td><td>37B</td></tr><tr><td></td><td># Total Params</td><td>-</td><td>-</td><td>671B</td><td>-</td><td>-</td><td>671B</td></tr><tr><td>English</td><td>MMLU (Pass@1)</td><td>88.3</td><td>87.2</td><td>88.5</td><td>85.2</td><td><strong>91.8</strong></td><td>90.8</td></tr><tr><td></td><td>MMLU-Redux (EM)</td><td>88.9</td><td>88.0</td><td>89.1</td><td>86.7</td><td>-</td><td><strong>92.9</strong></td></tr><tr><td></td><td>MMLU-Pro (EM)</td><td>78.0</td><td>72.6</td><td>75.9</td><td>80.3</td><td>-</td><td><strong>84.0</strong></td></tr><tr><td></td><td>DROP (3-shot F1)</td><td>88.3</td><td>83.7</td><td>91.6</td><td>83.9</td><td>90.2</td><td><strong>92.2</strong></td></tr><tr><td></td><td>IF-Eval (Prompt Strict)</td><td><strong>86.5</strong></td><td>84.3</td><td>86.1</td><td>84.8</td><td>-</td><td>83.3</td></tr><tr><td></td><td>GPQA-Diamond (Pass@1)</td><td>65.0</td><td>49.9</td><td>59.1</td><td>60.0</td><td><strong>75.7</strong></td><td>71.5</td></tr><tr><td></td><td>SimpleQA (Correct)</td><td>28.4</td><td>38.2</td><td>24.9</td><td>7.0</td><td><strong>47.0</strong></td><td>30.1</td></tr><tr><td></td><td>FRAMES (Acc.)</td><td>72.5</td><td>80.5</td><td>73.3</td><td>76.9</td><td>-</td><td><strong>82.5</strong></td></tr><tr><td></td><td>AlpacaEval2.0 (LC-winrate)</td><td>52.0</td><td>51.1</td><td>70.0</td><td>57.8</td><td>-</td><td><strong>87.6</strong></td></tr><tr><td></td><td>ArenaHard (GPT-4-1106)</td><td>85.2</td><td>80.4</td><td>85.5</td><td>92.0</td><td>-</td><td><strong>92.3</strong></td></tr><tr><td>Code</td><td>LiveCodeBench (Pass@1-COT)</td><td>33.8</td><td>34.2</td><td>-</td><td>53.8</td><td>63.4</td><td><strong>65.9</strong></td></tr><tr><td></td><td>Codeforces (Percentile)</td><td>20.3</td><td>23.6</td><td>58.7</td><td>93.4</td><td><strong>96.6</strong></td><td>96.3</td></tr><tr><td></td><td>Codeforces (Rating)</td><td>717</td><td>759</td><td>1134</td><td>1820</td><td><strong>2061</strong></td><td>2029</td></tr><tr><td></td><td>SWE Verified (Resolved)</td><td><strong>50.8</strong></td><td>38.8</td><td>42.0</td><td>41.6</td><td>48.9</td><td>49.2</td></tr><tr><td></td><td>Aider-Polyglot (Acc.)</td><td>45.3</td><td>16.0</td><td>49.6</td><td>32.9</td><td><strong>61.7</strong></td><td>53.3</td></tr><tr><td>Math</td><td>AIME 2024 (Pass@1)</td><td>16.0</td><td>9.3</td><td>39.2</td><td>63.6</td><td>79.2</td><td><strong>79.8</strong></td></tr><tr><td></td><td>MATH-500 (Pass@1)</td><td>78.3</td><td>74.6</td><td>90.2</td><td>90.0</td><td>96.4</td><td><strong>97.3</strong></td></tr><tr><td></td><td>CNMO 2024 (Pass@1)</td><td>13.1</td><td>10.8</td><td>43.2</td><td>67.6</td><td>-</td><td><strong>78.8</strong></td></tr><tr><td>Chinese</td><td>CLUEWSC (EM)</td><td>85.4</td><td>87.9</td><td>90.9</td><td>89.9</td><td>-</td><td><strong>92.8</strong></td></tr><tr><td></td><td>C-Eval (EM)</td><td>76.7</td><td>76.0</td><td>86.5</td><td>68.9</td><td>-</td><td><strong>91.8</strong></td></tr><tr><td></td><td>C-SimpleQA (Correct)</td><td>55.4</td><td>58.7</td><td><strong>68.0</strong></td><td>40.3</td><td>-</td><td>63.7</td></tr></table><p></p><p>DeepSeek R1 release <a href="https://api-docs.deepseek.com/news/news250120" target="_blank">https://api-docs.deepseek.com/news/news250120</a></p><p>- Online chat <a href="https://chat.deepseek.com/" target="_blank">https://chat.deepseek.com/</a></p><p>- Deepseek coder <a href="https://huggingface.co/spaces/akhaliq/anychat" target="_blank">https://huggingface.co/spaces/akhaliq/anychat</a></p><p>- Higgingface <a href="https://huggingface.co/deepseek-ai/DeepSeek-R1" target="_blank">https://huggingface.co/deepseek-ai/DeepSeek-R1</a></p><p></p><p><strong>DeepSeek Math Capabilities: Precision Meets Power</strong></p><p>DeepSeek’s math capabilities are designed to handle a wide range of mathematical tasks, from basic arithmetic to advanced calculus, linear algebra, and statistical analysis. Its AI-driven engine ensures accurate problem-solving, step-by-step explanations, and the ability to tackle complex equations with ease. Whether you're a student, educator, or professional, DeepSeek can simplify math challenges and provide clear, logical solutions.</p></blockquote><p></p>
[QUOTE="MihiCherub, post: 30496402, member: 238676"] The DeepSeek R1 model is a versatile, powerful AI tool that’s set to revolutionize how businesses and individuals interact with technology. Whether you’re automating workflows or enhancing user experiences, the R1 is a game-changer. [B]Open-Source Model You Can Run Locally[/B] [IMG width="771px"]https://dp-cdn-deepseek.obs.cn-east-3.myhuaweicloud.com/api-docs/deepseek_r1_example_en.gif[/IMG] [IMG width="770px"]https://huggingface.co/deepseek-ai/DeepSeek-R1/resolve/main/figures/benchmark.jpg[/IMG] [HEADING=2][/HEADING] [HEADING=2]DeepSeek-R1-Evaluation [/HEADING] [TABLE] [TR] [TH]Category[/TH] [TH]Benchmark (Metric)[/TH] [TH]Claude-3.5-Sonnet-1022[/TH] [TH]GPT-4o 0513[/TH] [TH]DeepSeek V3[/TH] [TH]OpenAI o1-mini[/TH] [TH]OpenAI o1-1217[/TH] [TH]DeepSeek R1[/TH] [/TR] [TR] [TD][/TD] [TD]Architecture[/TD] [TD]-[/TD] [TD]-[/TD] [TD]MoE[/TD] [TD]-[/TD] [TD]-[/TD] [TD]MoE[/TD] [/TR] [TR] [TD][/TD] [TD]# Activated Params[/TD] [TD]-[/TD] [TD]-[/TD] [TD]37B[/TD] [TD]-[/TD] [TD]-[/TD] [TD]37B[/TD] [/TR] [TR] [TD][/TD] [TD]# Total Params[/TD] [TD]-[/TD] [TD]-[/TD] [TD]671B[/TD] [TD]-[/TD] [TD]-[/TD] [TD]671B[/TD] [/TR] [TR] [TD]English[/TD] [TD]MMLU (Pass@1)[/TD] [TD]88.3[/TD] [TD]87.2[/TD] [TD]88.5[/TD] [TD]85.2[/TD] [TD][B]91.8[/B][/TD] [TD]90.8[/TD] [/TR] [TR] [TD][/TD] [TD]MMLU-Redux (EM)[/TD] [TD]88.9[/TD] [TD]88.0[/TD] [TD]89.1[/TD] [TD]86.7[/TD] [TD]-[/TD] [TD][B]92.9[/B][/TD] [/TR] [TR] [TD][/TD] [TD]MMLU-Pro (EM)[/TD] [TD]78.0[/TD] [TD]72.6[/TD] [TD]75.9[/TD] [TD]80.3[/TD] [TD]-[/TD] [TD][B]84.0[/B][/TD] [/TR] [TR] [TD][/TD] [TD]DROP (3-shot F1)[/TD] [TD]88.3[/TD] [TD]83.7[/TD] [TD]91.6[/TD] [TD]83.9[/TD] [TD]90.2[/TD] [TD][B]92.2[/B][/TD] [/TR] [TR] [TD][/TD] [TD]IF-Eval (Prompt Strict)[/TD] [TD][B]86.5[/B][/TD] [TD]84.3[/TD] [TD]86.1[/TD] [TD]84.8[/TD] [TD]-[/TD] [TD]83.3[/TD] [/TR] [TR] [TD][/TD] [TD]GPQA-Diamond (Pass@1)[/TD] [TD]65.0[/TD] [TD]49.9[/TD] [TD]59.1[/TD] [TD]60.0[/TD] [TD][B]75.7[/B][/TD] [TD]71.5[/TD] [/TR] [TR] [TD][/TD] [TD]SimpleQA (Correct)[/TD] [TD]28.4[/TD] [TD]38.2[/TD] [TD]24.9[/TD] [TD]7.0[/TD] [TD][B]47.0[/B][/TD] [TD]30.1[/TD] [/TR] [TR] [TD][/TD] [TD]FRAMES (Acc.)[/TD] [TD]72.5[/TD] [TD]80.5[/TD] [TD]73.3[/TD] [TD]76.9[/TD] [TD]-[/TD] [TD][B]82.5[/B][/TD] [/TR] [TR] [TD][/TD] [TD]AlpacaEval2.0 (LC-winrate)[/TD] [TD]52.0[/TD] [TD]51.1[/TD] [TD]70.0[/TD] [TD]57.8[/TD] [TD]-[/TD] [TD][B]87.6[/B][/TD] [/TR] [TR] [TD][/TD] [TD]ArenaHard (GPT-4-1106)[/TD] [TD]85.2[/TD] [TD]80.4[/TD] [TD]85.5[/TD] [TD]92.0[/TD] [TD]-[/TD] [TD][B]92.3[/B][/TD] [/TR] [TR] [TD]Code[/TD] [TD]LiveCodeBench (Pass@1-COT)[/TD] [TD]33.8[/TD] [TD]34.2[/TD] [TD]-[/TD] [TD]53.8[/TD] [TD]63.4[/TD] [TD][B]65.9[/B][/TD] [/TR] [TR] [TD][/TD] [TD]Codeforces (Percentile)[/TD] [TD]20.3[/TD] [TD]23.6[/TD] [TD]58.7[/TD] [TD]93.4[/TD] [TD][B]96.6[/B][/TD] [TD]96.3[/TD] [/TR] [TR] [TD][/TD] [TD]Codeforces (Rating)[/TD] [TD]717[/TD] [TD]759[/TD] [TD]1134[/TD] [TD]1820[/TD] [TD][B]2061[/B][/TD] [TD]2029[/TD] [/TR] [TR] [TD][/TD] [TD]SWE Verified (Resolved)[/TD] [TD][B]50.8[/B][/TD] [TD]38.8[/TD] [TD]42.0[/TD] [TD]41.6[/TD] [TD]48.9[/TD] [TD]49.2[/TD] [/TR] [TR] [TD][/TD] [TD]Aider-Polyglot (Acc.)[/TD] [TD]45.3[/TD] [TD]16.0[/TD] [TD]49.6[/TD] [TD]32.9[/TD] [TD][B]61.7[/B][/TD] [TD]53.3[/TD] [/TR] [TR] [TD]Math[/TD] [TD]AIME 2024 (Pass@1)[/TD] [TD]16.0[/TD] [TD]9.3[/TD] [TD]39.2[/TD] [TD]63.6[/TD] [TD]79.2[/TD] [TD][B]79.8[/B][/TD] [/TR] [TR] [TD][/TD] [TD]MATH-500 (Pass@1)[/TD] [TD]78.3[/TD] [TD]74.6[/TD] [TD]90.2[/TD] [TD]90.0[/TD] [TD]96.4[/TD] [TD][B]97.3[/B][/TD] [/TR] [TR] [TD][/TD] [TD]CNMO 2024 (Pass@1)[/TD] [TD]13.1[/TD] [TD]10.8[/TD] [TD]43.2[/TD] [TD]67.6[/TD] [TD]-[/TD] [TD][B]78.8[/B][/TD] [/TR] [TR] [TD]Chinese[/TD] [TD]CLUEWSC (EM)[/TD] [TD]85.4[/TD] [TD]87.9[/TD] [TD]90.9[/TD] [TD]89.9[/TD] [TD]-[/TD] [TD][B]92.8[/B][/TD] [/TR] [TR] [TD][/TD] [TD]C-Eval (EM)[/TD] [TD]76.7[/TD] [TD]76.0[/TD] [TD]86.5[/TD] [TD]68.9[/TD] [TD]-[/TD] [TD][B]91.8[/B][/TD] [/TR] [TR] [TD][/TD] [TD]C-SimpleQA (Correct)[/TD] [TD]55.4[/TD] [TD]58.7[/TD] [TD][B]68.0[/B][/TD] [TD]40.3[/TD] [TD]-[/TD] [TD]63.7[/TD] [/TR] [/TABLE] DeepSeek R1 release [URL]https://api-docs.deepseek.com/news/news250120[/URL] - Online chat [URL]https://chat.deepseek.com/[/URL] - Deepseek coder [URL]https://huggingface.co/spaces/akhaliq/anychat[/URL] - Higgingface [URL]https://huggingface.co/deepseek-ai/DeepSeek-R1[/URL] [B]DeepSeek Math Capabilities: Precision Meets Power[/B] DeepSeek’s math capabilities are designed to handle a wide range of mathematical tasks, from basic arithmetic to advanced calculus, linear algebra, and statistical analysis. Its AI-driven engine ensures accurate problem-solving, step-by-step explanations, and the ability to tackle complex equations with ease. Whether you're a student, educator, or professional, DeepSeek can simplify math challenges and provide clear, logical solutions. [/QUOTE]
Insert quotes…
Verification
Asuwa dahayen wadi kalama keeyada?
Post reply
Top
Bottom