[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-together-ai-together-evaluations":3,"mdc-rs41y0-key":34,"related-org-together-ai-together-evaluations":613,"related-repo-together-ai-together-evaluations":786},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"together-evaluations","evaluate LLM outputs with Together AI","LLM-as-a-judge evaluation framework on Together AI. Classify, score, and compare model outputs, select judge models, use external-provider judges or targets, poll results and download reports. Reach for it whenever the user wants to benchmark outputs, grade responses, compare A\u002FB variants, or operationalize automated evaluations.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},"together-ai","Together AI","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Ftogether-ai.jpg","togethercomputer",[13,17,20],{"name":14,"slug":15,"type":16},"Benchmarking","benchmarking","tag",{"name":18,"slug":19,"type":16},"LLM","llm",{"name":21,"slug":22,"type":16},"Evals","evals",31,"https:\u002F\u002Fgithub.com\u002Ftogethercomputer\u002Fskills","2026-07-26T05:49:07.241553",null,4,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"Skills to help your coding agents use Together AI products.","https:\u002F\u002Fgithub.com\u002Ftogethercomputer\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Ftogether-evaluations","---\nname: together-evaluations\ndescription: \"LLM-as-a-judge evaluation framework on Together AI. Classify, score, and compare model outputs, select judge models, use external-provider judges or targets, poll results and download reports. Reach for it whenever the user wants to benchmark outputs, grade responses, compare A\u002FB variants, or operationalize automated evaluations.\"\n---\n\n# Together AI Evaluations\n\n## Overview\n\nUse Together AI evaluations when the user wants a managed LLM-as-a-judge workflow rather than an\nad hoc prompt loop.\n\nCore evaluation types:\n\n- **Classify**: assign outputs to labels\n- **Score**: grade outputs on a numeric scale\n- **Compare**: compare two candidate outputs with bias controls\n\nThis skill also covers external providers used as judges or targets when the workflow still runs\nthrough Together AI's evaluation system.\n\n## When This Skill Wins\n\n- Benchmark prompt variants, models, or product responses\n- Grade quality, safety, policy compliance, or task success\n- Run A\u002FB comparisons between model outputs\n- Build repeatable evaluation jobs with uploaded datasets\n- Pull results programmatically after asynchronous execution\n\n## Hand Off To Another Skill\n\n- Use `together-chat-completions` for one-off inference or manual judge prompts\n- Use `together-batch-inference` for bulk offline generation rather than evaluation\n- Use `together-fine-tuning` when the user wants to improve the model instead of just measure it\n- Use `together-dedicated-model-inference` only if the evaluation target itself is a dedicated endpoint\n\n## Quick Routing\n\n- **Classify \u002F Score \u002F Compare job setup**\n  - Start with [scripts\u002Frun_evaluation.py](scripts\u002Frun_evaluation.py) or [scripts\u002Frun_evaluation.ts](scripts\u002Frun_evaluation.ts)\n  - Read [references\u002Fapi-reference.md](references\u002Fapi-reference.md) for exact request shapes\n- **Dataset formatting**\n  - Read the dataset sections in [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n- **Dataset columns, Jinja2 templates, or pre-generated responses**\n  - Read the dataset and template sections in [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n  - Use `--eval-column`, `--model-a-column`, or `--model-b-column` in the scripts\n- **Evaluating vision-capable models with image inputs**\n  - Read the Image Inputs section in [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n  - Add an `image_data_urls` column (base64 data URLs) to the dataset; the model and judge must be vision-capable\n- **External providers as judge or target**\n  - Read the model-source and provider sections in [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n  - Use the scripts with `--judge-model-source external`, `--eval-model-source external`, or compare-side source flags\n- **Polling, listing, or downloading results**\n  - Use the retrieval endpoints documented in [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n  - Use `--download-results` in the scripts when you want the per-row JSONL locally\n\n## Workflow\n\n1. Identify whether the user needs classify, score, or compare.\n2. Define the dataset schema before writing code.\n3. Upload the dataset as an eval file and keep the returned file ID.\n4. Configure judge and target models explicitly, especially when mixing providers.\n5. Poll status until completion, then download the result file for analysis.\n\n## High-Signal Rules\n\n- Python scripts require the Together v2 SDK (`together>=2.0.0`). If the user is on an older version, they must upgrade first: `uv pip install --upgrade \"together>=2.0.0\"`.\n- The current SDK examples in this repo use `check=False` for eval uploads because local file validation can misclassify eval datasets.\n- Treat dataset schema as part of the product contract; inconsistent fields cause downstream confusion.\n- Compare evaluations are best when both candidate responses are already present in the dataset.\n- Keep judge configuration explicit. Hidden defaults make benchmark interpretation harder.\n- Use Together AI's managed evaluation job instead of rebuilding a manual judge loop when repeatability matters.\n\n## Resource Map\n\n- **Full API reference**: [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n- **Dataset formats, Jinja2 templates, and provider shortcuts**: [references\u002Fapi-reference.md](references\u002Fapi-reference.md)\n- **Python workflow**: [scripts\u002Frun_evaluation.py](scripts\u002Frun_evaluation.py)\n- **TypeScript workflow**: [scripts\u002Frun_evaluation.ts](scripts\u002Frun_evaluation.ts)\n\n## Official Docs\n\n- [AI Evaluations](https:\u002F\u002Fdocs.together.ai\u002Fdocs\u002Fai-evaluations)\n- [Evaluations API](https:\u002F\u002Fdocs.together.ai\u002Freference\u002Fcreate-evaluation)\n",{"data":35,"body":36},{"name":4,"description":6},{"type":37,"children":38},"root",[39,48,55,61,66,102,107,113,141,147,200,206,423,429,458,464,521,527,583,589],{"type":40,"tag":41,"props":42,"children":44},"element","h1",{"id":43},"together-ai-evaluations",[45],{"type":46,"value":47},"text","Together AI Evaluations",{"type":40,"tag":49,"props":50,"children":52},"h2",{"id":51},"overview",[53],{"type":46,"value":54},"Overview",{"type":40,"tag":56,"props":57,"children":58},"p",{},[59],{"type":46,"value":60},"Use Together AI evaluations when the user wants a managed LLM-as-a-judge workflow rather than an\nad hoc prompt loop.",{"type":40,"tag":56,"props":62,"children":63},{},[64],{"type":46,"value":65},"Core evaluation types:",{"type":40,"tag":67,"props":68,"children":69},"ul",{},[70,82,92],{"type":40,"tag":71,"props":72,"children":73},"li",{},[74,80],{"type":40,"tag":75,"props":76,"children":77},"strong",{},[78],{"type":46,"value":79},"Classify",{"type":46,"value":81},": assign outputs to labels",{"type":40,"tag":71,"props":83,"children":84},{},[85,90],{"type":40,"tag":75,"props":86,"children":87},{},[88],{"type":46,"value":89},"Score",{"type":46,"value":91},": grade outputs on a numeric scale",{"type":40,"tag":71,"props":93,"children":94},{},[95,100],{"type":40,"tag":75,"props":96,"children":97},{},[98],{"type":46,"value":99},"Compare",{"type":46,"value":101},": compare two candidate outputs with bias controls",{"type":40,"tag":56,"props":103,"children":104},{},[105],{"type":46,"value":106},"This skill also covers external providers used as judges or targets when the workflow still runs\nthrough Together AI's evaluation system.",{"type":40,"tag":49,"props":108,"children":110},{"id":109},"when-this-skill-wins",[111],{"type":46,"value":112},"When This Skill Wins",{"type":40,"tag":67,"props":114,"children":115},{},[116,121,126,131,136],{"type":40,"tag":71,"props":117,"children":118},{},[119],{"type":46,"value":120},"Benchmark prompt variants, models, or product responses",{"type":40,"tag":71,"props":122,"children":123},{},[124],{"type":46,"value":125},"Grade quality, safety, policy compliance, or task success",{"type":40,"tag":71,"props":127,"children":128},{},[129],{"type":46,"value":130},"Run A\u002FB comparisons between model outputs",{"type":40,"tag":71,"props":132,"children":133},{},[134],{"type":46,"value":135},"Build repeatable evaluation jobs with uploaded datasets",{"type":40,"tag":71,"props":137,"children":138},{},[139],{"type":46,"value":140},"Pull results programmatically after asynchronous execution",{"type":40,"tag":49,"props":142,"children":144},{"id":143},"hand-off-to-another-skill",[145],{"type":46,"value":146},"Hand Off To Another Skill",{"type":40,"tag":67,"props":148,"children":149},{},[150,164,176,188],{"type":40,"tag":71,"props":151,"children":152},{},[153,155,162],{"type":46,"value":154},"Use ",{"type":40,"tag":156,"props":157,"children":159},"code",{"className":158},[],[160],{"type":46,"value":161},"together-chat-completions",{"type":46,"value":163}," for one-off inference or manual judge prompts",{"type":40,"tag":71,"props":165,"children":166},{},[167,168,174],{"type":46,"value":154},{"type":40,"tag":156,"props":169,"children":171},{"className":170},[],[172],{"type":46,"value":173},"together-batch-inference",{"type":46,"value":175}," for bulk offline generation rather than evaluation",{"type":40,"tag":71,"props":177,"children":178},{},[179,180,186],{"type":46,"value":154},{"type":40,"tag":156,"props":181,"children":183},{"className":182},[],[184],{"type":46,"value":185},"together-fine-tuning",{"type":46,"value":187}," when the user wants to improve the model instead of just measure it",{"type":40,"tag":71,"props":189,"children":190},{},[191,192,198],{"type":46,"value":154},{"type":40,"tag":156,"props":193,"children":195},{"className":194},[],[196],{"type":46,"value":197},"together-dedicated-model-inference",{"type":46,"value":199}," only if the evaluation target itself is a dedicated endpoint",{"type":40,"tag":49,"props":201,"children":203},{"id":202},"quick-routing",[204],{"type":46,"value":205},"Quick Routing",{"type":40,"tag":67,"props":207,"children":208},{},[209,250,270,318,351,391],{"type":40,"tag":71,"props":210,"children":211},{},[212,217],{"type":40,"tag":75,"props":213,"children":214},{},[215],{"type":46,"value":216},"Classify \u002F Score \u002F Compare job setup",{"type":40,"tag":67,"props":218,"children":219},{},[220,238],{"type":40,"tag":71,"props":221,"children":222},{},[223,225,231,233],{"type":46,"value":224},"Start with ",{"type":40,"tag":226,"props":227,"children":229},"a",{"href":228},"scripts\u002Frun_evaluation.py",[230],{"type":46,"value":228},{"type":46,"value":232}," or ",{"type":40,"tag":226,"props":234,"children":236},{"href":235},"scripts\u002Frun_evaluation.ts",[237],{"type":46,"value":235},{"type":40,"tag":71,"props":239,"children":240},{},[241,243,248],{"type":46,"value":242},"Read ",{"type":40,"tag":226,"props":244,"children":246},{"href":245},"references\u002Fapi-reference.md",[247],{"type":46,"value":245},{"type":46,"value":249}," for exact request shapes",{"type":40,"tag":71,"props":251,"children":252},{},[253,258],{"type":40,"tag":75,"props":254,"children":255},{},[256],{"type":46,"value":257},"Dataset formatting",{"type":40,"tag":67,"props":259,"children":260},{},[261],{"type":40,"tag":71,"props":262,"children":263},{},[264,266],{"type":46,"value":265},"Read the dataset sections in ",{"type":40,"tag":226,"props":267,"children":268},{"href":245},[269],{"type":46,"value":245},{"type":40,"tag":71,"props":271,"children":272},{},[273,278],{"type":40,"tag":75,"props":274,"children":275},{},[276],{"type":46,"value":277},"Dataset columns, Jinja2 templates, or pre-generated responses",{"type":40,"tag":67,"props":279,"children":280},{},[281,290],{"type":40,"tag":71,"props":282,"children":283},{},[284,286],{"type":46,"value":285},"Read the dataset and template sections in ",{"type":40,"tag":226,"props":287,"children":288},{"href":245},[289],{"type":46,"value":245},{"type":40,"tag":71,"props":291,"children":292},{},[293,294,300,302,308,310,316],{"type":46,"value":154},{"type":40,"tag":156,"props":295,"children":297},{"className":296},[],[298],{"type":46,"value":299},"--eval-column",{"type":46,"value":301},", ",{"type":40,"tag":156,"props":303,"children":305},{"className":304},[],[306],{"type":46,"value":307},"--model-a-column",{"type":46,"value":309},", or ",{"type":40,"tag":156,"props":311,"children":313},{"className":312},[],[314],{"type":46,"value":315},"--model-b-column",{"type":46,"value":317}," in the scripts",{"type":40,"tag":71,"props":319,"children":320},{},[321,326],{"type":40,"tag":75,"props":322,"children":323},{},[324],{"type":46,"value":325},"Evaluating vision-capable models with image inputs",{"type":40,"tag":67,"props":327,"children":328},{},[329,338],{"type":40,"tag":71,"props":330,"children":331},{},[332,334],{"type":46,"value":333},"Read the Image Inputs section in ",{"type":40,"tag":226,"props":335,"children":336},{"href":245},[337],{"type":46,"value":245},{"type":40,"tag":71,"props":339,"children":340},{},[341,343,349],{"type":46,"value":342},"Add an ",{"type":40,"tag":156,"props":344,"children":346},{"className":345},[],[347],{"type":46,"value":348},"image_data_urls",{"type":46,"value":350}," column (base64 data URLs) to the dataset; the model and judge must be vision-capable",{"type":40,"tag":71,"props":352,"children":353},{},[354,359],{"type":40,"tag":75,"props":355,"children":356},{},[357],{"type":46,"value":358},"External providers as judge or target",{"type":40,"tag":67,"props":360,"children":361},{},[362,371],{"type":40,"tag":71,"props":363,"children":364},{},[365,367],{"type":46,"value":366},"Read the model-source and provider sections in ",{"type":40,"tag":226,"props":368,"children":369},{"href":245},[370],{"type":46,"value":245},{"type":40,"tag":71,"props":372,"children":373},{},[374,376,382,383,389],{"type":46,"value":375},"Use the scripts with ",{"type":40,"tag":156,"props":377,"children":379},{"className":378},[],[380],{"type":46,"value":381},"--judge-model-source external",{"type":46,"value":301},{"type":40,"tag":156,"props":384,"children":386},{"className":385},[],[387],{"type":46,"value":388},"--eval-model-source external",{"type":46,"value":390},", or compare-side source flags",{"type":40,"tag":71,"props":392,"children":393},{},[394,399],{"type":40,"tag":75,"props":395,"children":396},{},[397],{"type":46,"value":398},"Polling, listing, or downloading results",{"type":40,"tag":67,"props":400,"children":401},{},[402,411],{"type":40,"tag":71,"props":403,"children":404},{},[405,407],{"type":46,"value":406},"Use the retrieval endpoints documented in ",{"type":40,"tag":226,"props":408,"children":409},{"href":245},[410],{"type":46,"value":245},{"type":40,"tag":71,"props":412,"children":413},{},[414,415,421],{"type":46,"value":154},{"type":40,"tag":156,"props":416,"children":418},{"className":417},[],[419],{"type":46,"value":420},"--download-results",{"type":46,"value":422}," in the scripts when you want the per-row JSONL locally",{"type":40,"tag":49,"props":424,"children":426},{"id":425},"workflow",[427],{"type":46,"value":428},"Workflow",{"type":40,"tag":430,"props":431,"children":432},"ol",{},[433,438,443,448,453],{"type":40,"tag":71,"props":434,"children":435},{},[436],{"type":46,"value":437},"Identify whether the user needs classify, score, or compare.",{"type":40,"tag":71,"props":439,"children":440},{},[441],{"type":46,"value":442},"Define the dataset schema before writing code.",{"type":40,"tag":71,"props":444,"children":445},{},[446],{"type":46,"value":447},"Upload the dataset as an eval file and keep the returned file ID.",{"type":40,"tag":71,"props":449,"children":450},{},[451],{"type":46,"value":452},"Configure judge and target models explicitly, especially when mixing providers.",{"type":40,"tag":71,"props":454,"children":455},{},[456],{"type":46,"value":457},"Poll status until completion, then download the result file for analysis.",{"type":40,"tag":49,"props":459,"children":461},{"id":460},"high-signal-rules",[462],{"type":46,"value":463},"High-Signal Rules",{"type":40,"tag":67,"props":465,"children":466},{},[467,488,501,506,511,516],{"type":40,"tag":71,"props":468,"children":469},{},[470,472,478,480,486],{"type":46,"value":471},"Python scripts require the Together v2 SDK (",{"type":40,"tag":156,"props":473,"children":475},{"className":474},[],[476],{"type":46,"value":477},"together>=2.0.0",{"type":46,"value":479},"). If the user is on an older version, they must upgrade first: ",{"type":40,"tag":156,"props":481,"children":483},{"className":482},[],[484],{"type":46,"value":485},"uv pip install --upgrade \"together>=2.0.0\"",{"type":46,"value":487},".",{"type":40,"tag":71,"props":489,"children":490},{},[491,493,499],{"type":46,"value":492},"The current SDK examples in this repo use ",{"type":40,"tag":156,"props":494,"children":496},{"className":495},[],[497],{"type":46,"value":498},"check=False",{"type":46,"value":500}," for eval uploads because local file validation can misclassify eval datasets.",{"type":40,"tag":71,"props":502,"children":503},{},[504],{"type":46,"value":505},"Treat dataset schema as part of the product contract; inconsistent fields cause downstream confusion.",{"type":40,"tag":71,"props":507,"children":508},{},[509],{"type":46,"value":510},"Compare evaluations are best when both candidate responses are already present in the dataset.",{"type":40,"tag":71,"props":512,"children":513},{},[514],{"type":46,"value":515},"Keep judge configuration explicit. Hidden defaults make benchmark interpretation harder.",{"type":40,"tag":71,"props":517,"children":518},{},[519],{"type":46,"value":520},"Use Together AI's managed evaluation job instead of rebuilding a manual judge loop when repeatability matters.",{"type":40,"tag":49,"props":522,"children":524},{"id":523},"resource-map",[525],{"type":46,"value":526},"Resource Map",{"type":40,"tag":67,"props":528,"children":529},{},[530,544,557,570],{"type":40,"tag":71,"props":531,"children":532},{},[533,538,540],{"type":40,"tag":75,"props":534,"children":535},{},[536],{"type":46,"value":537},"Full API reference",{"type":46,"value":539},": ",{"type":40,"tag":226,"props":541,"children":542},{"href":245},[543],{"type":46,"value":245},{"type":40,"tag":71,"props":545,"children":546},{},[547,552,553],{"type":40,"tag":75,"props":548,"children":549},{},[550],{"type":46,"value":551},"Dataset formats, Jinja2 templates, and provider shortcuts",{"type":46,"value":539},{"type":40,"tag":226,"props":554,"children":555},{"href":245},[556],{"type":46,"value":245},{"type":40,"tag":71,"props":558,"children":559},{},[560,565,566],{"type":40,"tag":75,"props":561,"children":562},{},[563],{"type":46,"value":564},"Python workflow",{"type":46,"value":539},{"type":40,"tag":226,"props":567,"children":568},{"href":228},[569],{"type":46,"value":228},{"type":40,"tag":71,"props":571,"children":572},{},[573,578,579],{"type":40,"tag":75,"props":574,"children":575},{},[576],{"type":46,"value":577},"TypeScript workflow",{"type":46,"value":539},{"type":40,"tag":226,"props":580,"children":581},{"href":235},[582],{"type":46,"value":235},{"type":40,"tag":49,"props":584,"children":586},{"id":585},"official-docs",[587],{"type":46,"value":588},"Official Docs",{"type":40,"tag":67,"props":590,"children":591},{},[592,603],{"type":40,"tag":71,"props":593,"children":594},{},[595],{"type":40,"tag":226,"props":596,"children":600},{"href":597,"rel":598},"https:\u002F\u002Fdocs.together.ai\u002Fdocs\u002Fai-evaluations",[599],"nofollow",[601],{"type":46,"value":602},"AI Evaluations",{"type":40,"tag":71,"props":604,"children":605},{},[606],{"type":40,"tag":226,"props":607,"children":610},{"href":608,"rel":609},"https:\u002F\u002Fdocs.together.ai\u002Freference\u002Fcreate-evaluation",[599],[611],{"type":46,"value":612},"Evaluations API",{"items":614,"total":785},[615,636,648,667,680,693,712,721,736,742,751,766],{"slug":616,"name":616,"fn":617,"description":618,"org":619,"tags":620,"stars":633,"repoUrl":634,"updatedAt":635},"blog-post","write structured blog posts","Write structured long-form blog posts with a consistent structure and SEO optimization.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[621,624,627,630],{"name":622,"slug":623,"type":16},"Content Creation","content-creation",{"name":625,"slug":626,"type":16},"Marketing","marketing",{"name":628,"slug":629,"type":16},"SEO","seo",{"name":631,"slug":632,"type":16},"Writing","writing",1151,"https:\u002F\u002Fgithub.com\u002Ftogethercomputer\u002Ftogether-cookbook","2026-07-17T06:07:39.475433",{"slug":637,"name":637,"fn":638,"description":639,"org":640,"tags":641,"stars":633,"repoUrl":634,"updatedAt":647},"social-media","create social media content","Create social media content including Twitter\u002FX threads, LinkedIn posts, and short-form updates.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[642,643,644,646],{"name":622,"slug":623,"type":16},{"name":625,"slug":626,"type":16},{"name":645,"slug":637,"type":16},"Social Media",{"name":631,"slug":632,"type":16},"2026-07-17T06:07:38.455058",{"slug":649,"name":649,"fn":650,"description":651,"org":652,"tags":653,"stars":23,"repoUrl":24,"updatedAt":666},"together-audio","process audio with Together AI","Text-to-speech and speech-to-text via Together AI, including REST, streaming, and realtime WebSocket TTS, plus transcription, translation, diarization, timestamps, and live STT. Reach for it whenever the user needs audio in or audio out on Together AI rather than chat generation, image or video creation, or model training.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[654,657,660,663],{"name":655,"slug":656,"type":16},"Audio","audio",{"name":658,"slug":659,"type":16},"Speech","speech",{"name":661,"slug":662,"type":16},"Text-to-Speech","text-to-speech",{"name":664,"slug":665,"type":16},"Transcription","transcription","2026-07-26T05:49:08.246858",{"slug":173,"name":173,"fn":668,"description":669,"org":670,"tags":671,"stars":23,"repoUrl":24,"updatedAt":679},"run asynchronous batch inference jobs","High-volume, asynchronous offline inference at up to 50% lower cost via Together AI's Batch API. Prepare JSONL inputs, upload files, create jobs, poll status, and download outputs. Reach for it whenever the user needs non-interactive bulk inference rather than real-time chat or evaluation jobs.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[672,675,676],{"name":673,"slug":674,"type":16},"Automation","automation",{"name":18,"slug":19,"type":16},{"name":677,"slug":678,"type":16},"Performance","performance","2026-07-17T06:08:23.919602",{"slug":161,"name":161,"fn":681,"description":682,"org":683,"tags":684,"stars":23,"repoUrl":24,"updatedAt":692},"generate text with Together AI","Real-time and streaming text generation via Together AI's OpenAI-compatible chat\u002Fcompletions API, including multi-turn conversations, tool and function calling, structured JSON outputs, and reasoning models. Reach for it whenever the user wants to build or debug text generation on Together AI, unless they specifically need batch jobs, embeddings, fine-tuning, dedicated endpoints, dedicated containers, or GPU clusters.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[685,688,691],{"name":686,"slug":687,"type":16},"AI","ai",{"name":689,"slug":690,"type":16},"API Development","api-development",{"name":18,"slug":19,"type":16},"2026-07-26T05:49:11.244104",{"slug":694,"name":694,"fn":695,"description":696,"org":697,"tags":698,"stars":23,"repoUrl":24,"updatedAt":711},"together-dedicated-containers","deploy custom inference containers on Together AI","Custom Dockerized inference workers on Together AI's managed GPU infrastructure. Build with Sprocket SDK, configure with Jig CLI, submit async queue jobs, and poll results. Reach for it whenever the user needs container-level control rather than a standard model endpoint or raw cluster.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[699,702,705,708],{"name":700,"slug":701,"type":16},"AI Infrastructure","ai-infrastructure",{"name":703,"slug":704,"type":16},"Deployment","deployment",{"name":706,"slug":707,"type":16},"Docker","docker",{"name":709,"slug":710,"type":16},"Machine Learning","machine-learning","2026-07-26T05:49:09.267892",{"slug":197,"name":197,"fn":713,"description":714,"org":715,"tags":716,"stars":23,"repoUrl":24,"updatedAt":720},"deploy and operate Together AI models","Deploy and operate models on dedicated GPUs with Together AI's Dedicated Model Inference (DMI, the v2 dedicated endpoints API): beta endpoints, deployments, deployment profiles and hardware configs, autoscaling, traffic splitting, A\u002FB tests, shadow experiments, Prometheus metrics, and custom model or LoRA adapter uploads. Reach for it whenever the user mentions together beta endpoints or tg beta commands, client.beta.endpoints, DMI resources like ep_\u002Fdep_\u002Fcr_\u002Fml_ IDs, or wants production model serving with traffic management on Together AI. This is the current dedicated-hosting API and also covers migrating off the retired legacy v1 endpoints API (non-beta client.endpoints \u002F together endpoints), whose create and restart now return HTTP 403.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[717,718,719],{"name":700,"slug":701,"type":16},{"name":703,"slug":704,"type":16},{"name":18,"slug":19,"type":16},"2026-07-26T06:08:44.044039",{"slug":722,"name":722,"fn":723,"description":724,"org":725,"tags":726,"stars":23,"repoUrl":24,"updatedAt":735},"together-embeddings","generate embeddings and build RAG pipelines","Dense vector embeddings, semantic search, RAG pipelines, and reranking via Together AI. Generate embeddings with open-source models and rerank results behind dedicated endpoints. Reach for it whenever the user needs vector representations or retrieval quality improvements rather than direct text generation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[727,728,729,732],{"name":686,"slug":687,"type":16},{"name":18,"slug":19,"type":16},{"name":730,"slug":731,"type":16},"RAG","rag",{"name":733,"slug":734,"type":16},"Search","search","2026-07-26T05:49:06.247906",{"slug":4,"name":4,"fn":5,"description":6,"org":737,"tags":738,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[739,740,741],{"name":14,"slug":15,"type":16},{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"slug":185,"name":185,"fn":743,"description":744,"org":745,"tags":746,"stars":23,"repoUrl":24,"updatedAt":750},"fine-tune and adapt models on Together AI","LoRA, full fine-tuning, DPO preference tuning, VLM training, function-calling tuning, reasoning tuning, and BYOM uploads on Together AI. Reach for it whenever the user wants to adapt a model on custom data rather than only run inference, evaluate outputs, or host an existing model.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[747,748,749],{"name":686,"slug":687,"type":16},{"name":18,"slug":19,"type":16},{"name":709,"slug":710,"type":16},"2026-07-26T05:49:10.243114",{"slug":752,"name":752,"fn":753,"description":754,"org":755,"tags":756,"stars":23,"repoUrl":24,"updatedAt":765},"together-gpu-clusters","orchestrate GPU clusters on Together AI","On-demand and reserved GPU clusters (H100, H200, B200) on Together AI with Kubernetes or Slurm orchestration, shared storage, credential management, and cluster scaling for ML and HPC jobs. Reach for it when the user needs multi-node compute or infrastructure control rather than a managed model endpoint.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[757,758,761,764],{"name":700,"slug":701,"type":16},{"name":759,"slug":760,"type":16},"Cloud","cloud",{"name":762,"slug":763,"type":16},"Kubernetes","kubernetes",{"name":709,"slug":710,"type":16},"2026-07-26T05:49:05.252646",{"slug":767,"name":767,"fn":768,"description":769,"org":770,"tags":771,"stars":23,"repoUrl":24,"updatedAt":784},"together-images","generate and edit images with Together AI","Text-to-image generation and image editing via Together AI, including FLUX and Kontext models, LoRA-based styling, reference-image guidance, and local image downloads. Reach for it whenever the user wants to generate or edit images on Together AI rather than create videos or build text-only chat applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[772,775,778,781],{"name":773,"slug":774,"type":16},"Creative","creative",{"name":776,"slug":777,"type":16},"Design","design",{"name":779,"slug":780,"type":16},"Image Generation","image-generation",{"name":782,"slug":783,"type":16},"Multimodal","multimodal","2026-07-17T06:04:23.067755",16,{"items":787,"total":833},[788,795,801,807,814,820,827],{"slug":649,"name":649,"fn":650,"description":651,"org":789,"tags":790,"stars":23,"repoUrl":24,"updatedAt":666},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[791,792,793,794],{"name":655,"slug":656,"type":16},{"name":658,"slug":659,"type":16},{"name":661,"slug":662,"type":16},{"name":664,"slug":665,"type":16},{"slug":173,"name":173,"fn":668,"description":669,"org":796,"tags":797,"stars":23,"repoUrl":24,"updatedAt":679},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[798,799,800],{"name":673,"slug":674,"type":16},{"name":18,"slug":19,"type":16},{"name":677,"slug":678,"type":16},{"slug":161,"name":161,"fn":681,"description":682,"org":802,"tags":803,"stars":23,"repoUrl":24,"updatedAt":692},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[804,805,806],{"name":686,"slug":687,"type":16},{"name":689,"slug":690,"type":16},{"name":18,"slug":19,"type":16},{"slug":694,"name":694,"fn":695,"description":696,"org":808,"tags":809,"stars":23,"repoUrl":24,"updatedAt":711},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[810,811,812,813],{"name":700,"slug":701,"type":16},{"name":703,"slug":704,"type":16},{"name":706,"slug":707,"type":16},{"name":709,"slug":710,"type":16},{"slug":197,"name":197,"fn":713,"description":714,"org":815,"tags":816,"stars":23,"repoUrl":24,"updatedAt":720},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[817,818,819],{"name":700,"slug":701,"type":16},{"name":703,"slug":704,"type":16},{"name":18,"slug":19,"type":16},{"slug":722,"name":722,"fn":723,"description":724,"org":821,"tags":822,"stars":23,"repoUrl":24,"updatedAt":735},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[823,824,825,826],{"name":686,"slug":687,"type":16},{"name":18,"slug":19,"type":16},{"name":730,"slug":731,"type":16},{"name":733,"slug":734,"type":16},{"slug":4,"name":4,"fn":5,"description":6,"org":828,"tags":829,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[830,831,832],{"name":14,"slug":15,"type":16},{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},14]