[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-llm-benchmark-2026-38-real-tasks-15-models-en":3,"article-related-llm-benchmark-2026-38-real-tasks-15-models-en":33,"series-industry-0eb521a5-8e5c-4825-884f-18f5f045bca7":76},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":25,"views":29,"created_at":30,"published_at":31,"topic_cluster_id":32},"0eb521a5-8e5c-4825-884f-18f5f045bca7","llm-benchmark-2026-38-real-tasks-15-models-en","15 LLMs on 38 tasks show routing beats one-model bets","\u003Cp>Which \u003Ca href=\"\u002Ftag\u002Fllm\">LLM\u003C\u002Fa> should you use for real work in 2026?\u003C\u002Fp>\u003Cp data-speakable=\"summary\">This \u003Ca href=\"\u002Ftag\u002Fbenchmark\">benchmark\u003C\u002Fa> compares 15 models on 38 real tasks to show which ones are worth routing to.\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>Item\u003C\u002Fth>\u003Cth>Quality\u003C\u002Fth>\u003Cth>Pass rate\u003C\u002Fth>\u003Cth>Median time\u003C\u002Fth>\u003Cth>Total cost\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>Claude Sonnet 4.6\u003C\u002Ftd>\u003Ctd>100%\u003C\u002Ftd>\u003Ctd>38\u002F38\u003C\u002Ftd>\u003Ctd>4.6s\u003C\u002Ftd>\u003Ctd>$0.20\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Claude Opus 4.6\u003C\u002Ftd>\u003Ctd>100%\u003C\u002Ftd>\u003Ctd>38\u002F38\u003C\u002Ftd>\u003Ctd>4.1s\u003C\u002Ftd>\u003Ctd>$0.69\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>MiniMax M2.5\u003C\u002Ftd>\u003Ctd>98.6%\u003C\u002Ftd>\u003Ctd>38\u002F38\u003C\u002Ftd>\u003Ctd>15.9s\u003C\u002Ftd>\u003Ctd>$0.07\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Gemini 2.5 Flash\u003C\u002Ftd>\u003Ctd>97.1%\u003C\u002Ftd>\u003Ctd>35\u002F38\u003C\u002Ftd>\u003Ctd>1.1s\u003C\u002Ftd>\u003Ctd>$0.003\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>GPT-oss-20b\u003C\u002Ftd>\u003Ctd>98.3%\u003C\u002Ftd>\u003Ctd>37\u002F38\u003C\u002Ftd>\u003Ctd>4.1s\u003C\u002Ftd>\u003Ctd>$0.00\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>1. Claude Sonnet 4.6\u003C\u002Fh2>\u003Cp>\u003Ca href=\"https:\u002F\u002Fwww.anthropic.com\u002Fclaude\">Claude Sonnet 4.6\u003C\u002Fa> is the cleanest all-around pick in the set: 100% pass rate, 4.6-second median time, and $0.20 total cost across the benchmark. It hit the ceiling on this test suite without needing the highest spend.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784422964229-9vo4.png\" alt=\"15 LLMs on 38 tasks show routing beats one-model bets\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>That combination matters because it shows you do not need to pay Opus pricing to get perfect results on these tasks. If your work mixes coding, extraction, and drafting, Sonnet is the safest default when you want one model to do most things well.\u003C\u002Fp>\u003Cul>\u003Cli>38\u002F38 tasks passed\u003C\u002Fli>\u003Cli>172.5\u002F172.5 points\u003C\u002Fli>\u003Cli>4.6s median response time\u003C\u002Fli>\u003Cli>$0.20 total benchmark cost\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>2. Claude Opus 4.6\u003C\u002Fh2>\u003Cp>\u003Ca href=\"https:\u002F\u002Fwww.anthropic.com\u002Fclaude\">Claude Opus 4.6\u003C\u002Fa> matched Sonnet on accuracy, but cost 3.5x more at $0.69 total. The benchmark does not show a quality edge that justifies the extra spend for these tasks.\u003C\u002Fp>\u003Cp>That makes Opus a specialist choice, not the first model to reach for. If you already know a task is hard, or if you want the extra margin for sensitive reasoning, Opus is still a strong option. But on this evidence, it is not the value winner.\u003C\u002Fp>\u003Cul>\u003Cli>38\u002F38 tasks passed\u003C\u002Fli>\u003Cli>4.1s median response time\u003C\u002Fli>\u003Cli>Same accuracy as Sonnet\u003C\u002Fli>\u003Cli>Highest cost among the top two\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>3. MiniMax M2.5\u003C\u002Fh2>\u003Cp>\u003Ca href=\"https:\u002F\u002Fwww.minimax.io\u002F\">MiniMax M2.5\u003C\u002Fa> is the format compliance pick. It scored 98.6% with a 100% pass rate and returned bare JSON on most tasks, which makes it unusually good for automation.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784422960882-tt51.png\" alt=\"15 LLMs on 38 tasks show routing beats one-model bets\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>The practical value here is not just accuracy, it is output discipline. Many models add wrapper text or markdown that breaks parsers; MiniMax mostly does not. If your pipeline depends on strict structure, this model deserves a close look.\u003C\u002Fp>\u003Cul>\u003Cli>98.6% quality\u003C\u002Fli>\u003Cli>38\u002F38 pass rate\u003C\u002Fli>\u003Cli>$0.07 total cost\u003C\u002Fli>\u003Cli>23 of 38 outputs were JSON-only\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>4. Gemini 2.5 Flash\u003C\u002Fh2>\u003Cp>\u003Ca href=\"https:\u002F\u002Fai.google.dev\u002Fgemini-api\">Gemini 2.5 Flash\u003C\u002Fa> defines the low-cost, low-latency end of the chart: 1.1-second median time and just $0.003 total cost. It still reached 97.1% quality, which is enough for a lot of batch and extraction work.\u003C\u002Fp>\u003Cp>This is the model to route to when speed and price matter more than squeezing out the last few points. It failed only on reasoning-adjacent tasks, which means it is a strong fit for data cleanup, transformation, and other predictable jobs.\u003C\u002Fp>\u003Cul>\u003Cli>97.1% quality\u003C\u002Fli>\u003Cli>35\u002F38 pass rate\u003C\u002Fli>\u003Cli>1.1s median response time\u003C\u002Fli>\u003Cli>$0.003 total cost\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>5. GPT-oss-20b\u003C\u002Fh2>\u003Cp>\u003Ca href=\"https:\u002F\u002Fopenai.com\u002Fresearch\u002Fopenai-oss\">GPT-oss-20b\u003C\u002Fa> is the free surprise. It scored 98.3% with a 97% pass rate while running locally, which puts it ahead of Haiku, R1, and GPT-5-Nano in this benchmark.\u003C\u002Fp>\u003Cp>That matters if you care about on-prem use or cost control. Free does not mean toy here, and the result shows that a local model can be good enough for real production work when the task is not deeply reasoning-heavy.\u003C\u002Fp>\u003Cul>\u003Cli>98.3% quality\u003C\u002Fli>\u003Cli>37\u002F38 pass rate\u003C\u002Fli>\u003Cli>$0.00 total cost\u003C\u002Fli>\u003Cli>Runs locally on-prem\u003C\u002Fli>\u003C\u002Ful>\u003Ch2>How to decide\u003C\u002Fh2>\u003Cp>If you want one model for most work, pick Sonnet. If you need the best structured-output behavior, MiniMax M2.5 is the one to test first. If you need the cheapest fast path for routine jobs, \u003Ca href=\"\u002Ftag\u002Fgemini\">Gemini\u003C\u002Fa> Flash is hard to beat. If your priority is zero marginal cost and local control, GPT-oss-20b is the standout.\u003C\u002Fp>\u003Cp>The deeper lesson is routing, not loyalty. This benchmark suggests most teams should send easy tasks to cheaper models and reserve the expensive ones for reasoning-heavy work, where the quality gap is real.\u003C\u002Fp>","15 models on 38 real tasks show when to route to Sonnet, Flash, MiniMax, or a free local model.","ianlpaterson.com","https:\u002F\u002Fianlpaterson.com\u002Fblog\u002Fllm-benchmark-2026-38-actual-tasks-15-models-for-2-29\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784422964229-9vo4.png","industry","en","0c3a0e19-e3c7-48ae-8f7b-198ca0911957",[17,18,19,20,21,22,23,24],"LLM benchmark","model routing","Claude Sonnet","MiniMax M2.5","Gemini Flash","GPT-oss-20b","real-world tasks","cost comparison",[26,27,28],"Sonnet is the best all-around default in this benchmark.","MiniMax M2.5 is the strongest structured-output choice.","Gemini Flash is the cheapest fast option for routine tasks.",1,"2026-07-19T01:02:19.445604+00:00","2026-07-19T01:02:19.436+00:00","de4e6983-4b7b-4e34-b50e-d849cecd2a35",{"tags":34,"relatedLang":35,"relatedPosts":39},[],{"id":15,"slug":36,"title":37,"language":38},"llm-routing-benchmark-38-tasks-15-models-zh","5 款 LLM 的實戰路由結論","zh",[40,46,52,58,64,70],{"id":41,"slug":42,"title":43,"cover_image":44,"image_url":44,"created_at":45,"category":13},"b262ef33-ea4b-43c4-85b3-679f1ab51a0c","low-power-chinese-chips-winning-smart-glasses-en","Low-Power Chinese Chips Are Winning Smart Glasses Before AR Displays …","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784491386900-5rm0.png","2026-07-19T20:02:41.03737+00:00",{"id":47,"slug":48,"title":49,"cover_image":50,"image_url":50,"created_at":51,"category":13},"0f47800d-8320-4cb7-9800-69f0bc354300","nvidia-open-gpu-docs-defensive-concession-en","Nvidia’s “Open” GPU Docs Are a Defensive Concession, Not a Real Openi…","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784489577340-yqsi.png","2026-07-19T19:32:31.130791+00:00",{"id":53,"slug":54,"title":55,"cover_image":56,"image_url":56,"created_at":57,"category":13},"b63a6ae7-f38d-4894-890c-6b160d1cd505","wgu-anthropic-ai-native-credentialing-model-en","WGU’s Anthropic deal points to AI-native credentials","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784487764491-ci06.png","2026-07-19T19:02:22.51788+00:00",{"id":59,"slug":60,"title":61,"cover_image":62,"image_url":62,"created_at":63,"category":13},"06e3fc0d-a757-4fa7-ba47-0d8b43faafec","agi-2026-five-facts-cut-through-noise-en","AGI in 2026: 5 facts that cut through the noise","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784440964515-89h3.png","2026-07-19T06:02:21.576068+00:00",{"id":65,"slug":66,"title":67,"cover_image":68,"image_url":68,"created_at":69,"category":13},"530266f1-c175-4a07-aa6c-963662096df6","openai-staff-fund-rival-super-pac-en","$215,000 from OpenAI staff backs anti-AI PAC","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784419367270-a4em.png","2026-07-19T00:02:21.711542+00:00",{"id":71,"slug":72,"title":73,"cover_image":74,"image_url":74,"created_at":75,"category":13},"4abfb46f-b476-4837-abe0-823deb9bef92","open-source-agent-stacks-seven-layers-2026-en","Open source agent stacks split into seven layers in 2026","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1784403178969-fbjv.png","2026-07-18T19:32:33.06165+00:00",[77,82,87,92,97,102,107,112,117,122],{"id":78,"slug":79,"title":80,"created_at":81},"d35a1bd9-e709-412e-a2df-392df1dc572a","ai-impact-2026-developments-market-en","AI's Impact in 2026: Key Developments and Market Shifts","2026-03-25T16:20:33.205823+00:00",{"id":83,"slug":84,"title":85,"created_at":86},"5ed27921-5fd6-492e-8c59-78393bf37710","trumps-ai-legislative-framework-en","Trump's AI Legislative Framework: What's Inside?","2026-03-25T16:22:20.005325+00:00",{"id":88,"slug":89,"title":90,"created_at":91},"e454a642-f03c-4794-b185-5f651aebbaca","nvidia-gtc-2026-key-highlights-innovations-en","NVIDIA GTC 2026: Key Highlights and Innovations","2026-03-25T16:22:47.882615+00:00",{"id":93,"slug":94,"title":95,"created_at":96},"0ebb5b16-774a-4922-945d-5f2ce1df5a6d","claude-usage-diversifies-learning-curves-en","Claude Usage Diversifies, Learning Curves Emerge","2026-03-25T16:25:50.770376+00:00",{"id":98,"slug":99,"title":100,"created_at":101},"69934e86-2fc5-4280-8223-7b917a48ace8","openclaw-ai-commoditization-concerns-en","OpenClaw's Rise Raises Concerns of AI Model Commoditization","2026-03-25T16:26:30.582047+00:00",{"id":103,"slug":104,"title":105,"created_at":106},"b4b2575b-2ac8-46b2-b90e-ab1d7c060797","google-gemini-ai-rollout-2026-en","Google's Gemini AI Rollout Extended to 2026","2026-03-25T16:28:14.808842+00:00",{"id":108,"slug":109,"title":110,"created_at":111},"6e18bc65-42ae-4ad0-b564-67d7f66b979e","meta-llama4-fabricated-results-scandal-en","Meta's Llama 4 Scandal: Fabricated AI Test Results Unveiled","2026-03-25T16:29:15.482836+00:00",{"id":113,"slug":114,"title":115,"created_at":116},"bf888e9d-08be-4f47-996c-7b24b5ab3500","accenture-mistral-ai-deployment-en","Accenture and Mistral AI Team Up for AI Deployment","2026-03-25T16:31:01.894655+00:00",{"id":118,"slug":119,"title":120,"created_at":121},"5382b536-fad2-49c6-ac85-9eb2bae49f35","mistral-ai-high-stakes-2026-en","Mistral AI: Facing High Stakes in 2026","2026-03-25T16:31:39.941974+00:00",{"id":123,"slug":124,"title":125,"created_at":126},"9da3d2d6-b669-4971-ba1d-17fdb3548ed5","cursors-meteoric-rise-pressures-en","Cursor's Meteoric Rise Faces Industry Pressures","2026-03-25T16:32:21.899217+00:00"]