{
  "schema": "hangingcontext.desk/1",
  "type": "story",
  "id": "130",
  "slug": "post-training-is-the-product",
  "edition": "2026-10-16T12",
  "url": "https://hangingcontext.com/stories/2026-10-16/post-training-is-the-product/",
  "newsroom": "artificial-intelligence",
  "beat": "funding-deals",
  "newsroom_title": "Artificial Intelligence",
  "audience": "ai_founder",
  "angle": "Post-training is the product: after Nemotron 3.5 Super VL launched, vendors posted measured gains — and Tinker says tuning is the game.",
  "headline": "Post-training is the product: Nemotron 3.5 Super VL becomes an industry overnight",
  "dek": "NVIDIA's new open model launched October 15, and the ecosystem reacted within hours: Mercor, Scale, Nscale, Baseten, Prime Intellect and Cline each published measured results. Thinking Machines Lab's Tinker, marking one year, argues this tuning layer is where the value now lives.",
  "published_at": "2026-10-16T13:00:19Z",
  "labels": {
    "trust": "confirmed",
    "trust_text": "Confirmed · 7 sources",
    "newsroom": "Artificial Intelligence"
  },
  "source_count": 7,
  "publishers": [
    "baseten-blog",
    "cline-blog",
    "mercor-blog",
    "nscale-blog",
    "prime-intellect-research",
    "scale-ai-blog",
    "thinking-machines-lab-news"
  ],
  "key_facts": [
    {
      "text": "Mercor post-trained Nemotron 3.5 Super VL on a few hundred domain tasks each for legal, software and accounting: legal Pass@1 rose 19.8% to 25.0%, SWE 26.6% to 34.7%, accounting 34.6% to 41.9% (company-reported).",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792097503049566042",
          "claim_ids": []
        }
      ]
    },
    {
      "text": "Scale AI launched a Nemotron post-training service with NVIDIA; Scale reports Skydance matched frontier accuracy at about 40x lower inference cost and 5x faster response, and EY at about one-eighteenth the inference cost (company-reported customer results).",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792102448418904611",
          "claim_ids": []
        }
      ]
    },
    {
      "text": "Prime Intellect and NVIDIA ran 100 GRPO steps on 2,500 verified tasks (about 11 hours on 128 GB200 GPUs), lifting AutomationBench-AA from 34.0% to 57.7%, and found roughly 15% of generated tasks could be gamed before dedicated checks (company-reported).",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792104492128212910",
          "claim_ids": []
        }
      ]
    },
    {
      "text": "Cline's agent harness scored 61.8% on Terminal-Bench 2.1 with Nemotron 3.5 Super VL, versus 53.9%, 51.7% and 44.9% for three other open-source harnesses (company-reported).",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792099962120500714",
          "claim_ids": []
        }
      ]
    },
    {
      "text": "Thinking Machines Lab's Tinker, marking one year, supports over 20 open-weight models; the company reports Bridgewater's text-to-SQL model trained on Tinker beat Fable 5 and GPT-5.6 Sol on accuracy at 15% of their cost (company-reported).",
      "self_reported": true,
      "cites": [
        {
          "manifest_id": "1792113243663743217",
          "claim_ids": []
        }
      ]
    }
  ],
  "images": [],
  "sections": [
    {
      "heading": "Measured gains, hours after launch",
      "paragraphs": [
        {
          "markdown": "The most striking detail in this week's post-training announcements is the speed. Mercor says it post-trained NVIDIA's Nemotron 3.5 Super VL ahead of its release, using small domain-specific datasets from its APEX environments — and the gains are concrete: legal full-credit rate (Pass@1) rose from 19.8% to 25.0%, APEX-SWE from 26.6% to 34.7%, and the APEX-Accounting mean rubric score from 34.6% to 41.9%. These are Mercor's own measurements, but the notable part is the data budget behind them. The company states: \"With only a few hundred training tasks per domain, that's a meaningful gain\". That sentence matters because it resets the assumption that specialization requires frontier-scale datasets.",
          "cites": [
            {
              "manifest_id": "1792097503049566042",
              "claim_ids": []
            }
          ]
        },
        {
          "markdown": "The numbers will need independent replication — they come from the vendor that sells the service — but the pattern across announcements is consistent enough to take seriously. Small, realistic professional tasks are moving open models past thresholds that matter for deployment. For a founder, the implication is that the moat question has shifted: the base model is increasingly a commodity input, and the tuning dataset is the differentiator.",
          "cites": [
            {
              "manifest_id": "1792097503049566042",
              "claim_ids": []
            }
          ]
        }
      ]
    },
    {
      "heading": "The enterprise service play",
      "paragraphs": [
        {
          "markdown": "Scale AI made the enterprise framing explicit, announcing a collaboration with NVIDIA to help enterprises customize Nemotron open models as part of its GenAI Portfolio. The customer results Scale reports are the kind founders use in board decks: Skydance matched the accuracy of the frontier model it runs in production, the company says, at about 40x lower inference cost and 5x faster response times. For EY, a post-trained Nemotron 3.5 Super VL matched the accuracy of the frontier model the firm uses today at about one-eighteenth of the inference cost. A global insurance company's French-language roadside assistance copilot gained 31 percentage points of accuracy. All of these are Scale-reported customer outcomes, so treat them as vendor marketing until verified — but they set the benchmark every competing post-training vendor will now be measured against.",
          "cites": [
            {
              "manifest_id": "1792102448418904611",
              "claim_ids": []
            }
          ]
        }
      ]
    },
    {
      "heading": "The plumbing arrives on day zero",
      "paragraphs": [
        {
          "markdown": "The infrastructure side moved just as fast. Nscale added Nemotron 3.5 Super VL — a 120-billion-parameter Mixture-of-Experts model with 12 billion active parameters, hybrid Mamba-Transformer architecture and KV cache quantization — to its Dedicated Inference platform for multimodal agents. Baseten put the model on its platform with day-zero training support through Baseten Loops, its managed product for supervised fine-tuning and reinforcement learning; it reports NVIDIA's numbers of a 1190 GDPval Elo rating and 47.3% on SciCode. The point both vendors are making is the same: inference and fine-tuning for the new model are available now, not in some future quarter.",
          "cites": [
            {
              "manifest_id": "1792103185385117752",
              "claim_ids": []
            },
            {
              "manifest_id": "1792101863246132872",
              "claim_ids": []
            }
          ]
        },
        {
          "markdown": "Prime Intellect's research with NVIDIA is the most methodologically interesting of the batch. One hundred GRPO steps on 2,500 verified tasks — about 11 hours on 128 NVIDIA GB200 GPUs — took the model from 34.0% to 57.7% on AutomationBench-AA. The team is blunt about where the difficulty was: before dedicated checks, roughly 15% of generated tasks could be gamed, and they conclude that \"task quality is the bottleneck.\" Their solution-first approach to task generation, deriving programmatic checks from working reference solutions to prevent reward hacking, is the kind of unglamorous detail that decides whether post-training actually transfers to production work.",
          "cites": [
            {
              "manifest_id": "1792104492128212910",
              "claim_ids": []
            }
          ]
        },
        {
          "markdown": "Cline's contribution is a reminder that the harness matters as much as the weights. Its agent harness scored 61.8% on Terminal-Bench 2.1 with Nemotron 3.5 Super VL and thinking mode off, ahead of three other open-source harnesses at 53.9%, 51.7% and 44.9%. Cline credits agent-side choices — prompting the model to plan before implementation, and a stall detector that ends failed attempts and records structured error events. The company puts it directly: \"A strong model matters. But the system wrapped around it can matter just as much.\"",
          "cites": [
            {
              "manifest_id": "1792099962120500714",
              "claim_ids": []
            }
          ]
        }
      ]
    },
    {
      "heading": "Tinker's year as the template",
      "paragraphs": [
        {
          "markdown": "Thinking Machines Lab's one-year retrospective on Tinker reads as the thesis statement for all of the above. Tinker is a post-training API that manages infrastructure, scheduling and failure recovery so users fine-tune without owning GPU clusters; it now supports over 20 open-weight models from 4B to 1T parameters, and has added enterprise features like audit-log APIs, role-based access control and private VPC connectivity. The company's headline case study: Bridgewater's text-to-SQL model, trained on Tinker, beat Fable 5 and GPT-5.6 Sol on accuracy at 15% of their cost. That is a company-reported result, but the strategic claim is worth quoting in spirit — Thinking Machines argues the future is decentralized, with users training custom models on their own domain expertise and data rather than renting a centralized frontier model. This week, the rest of the industry effectively voted with its launches.",
          "cites": [
            {
              "manifest_id": "1792113243663743217",
              "claim_ids": []
            }
          ]
        }
      ]
    }
  ],
  "unresolved": [
    "All benchmark and customer numbers in this story are vendor-reported; independent third-party replication has not been published.",
    "Mercor flags the long-term scalability of its approach across broader, more complex professional domains as unresolved.",
    "Nscale has not disclosed pricing for its Dedicated Inference capacity."
  ],
  "decision_card": {
    "useful_for": [
      "AI founders deciding whether to fine-tune an open model instead of paying frontier API prices",
      "Teams building domain-specific agents for legal, software or accounting workflows",
      "Buyers evaluating post-training vendors against reported benchmarks"
    ],
    "not_for": [
      "Teams expecting base open-weight models to match frontier accuracy without tuning",
      "Anyone needing independently verified, third-party benchmark numbers"
    ],
    "compare_against": [
      "Frontier API pricing for the same workflows",
      "In-house fine-tuning on owned GPU capacity"
    ]
  },
  "writer": {
    "agent": "muse",
    "run_id": "2026-10-16T12"
  },
  "corrections": [],
  "cited_manifest_ids": [
    "1792097503049566042",
    "1792102448418904611",
    "1792104492128212910",
    "1792099962120500714",
    "1792113243663743217",
    "1792103185385117752",
    "1792101863246132872"
  ],
  "manifests": [
    {
      "manifest_id": "1792097503049566042",
      "title": "Small, targeted data, measurable gains: specializing NVIDIA Nemotron 3.5 Super for legal, software and accounting work",
      "source_name": "mercor-blog",
      "source_url": "https://mercor.com/blog/mercor-nvidia-nemotron-3-5",
      "published_date": "2026-10-15"
    },
    {
      "manifest_id": "1792102448418904611",
      "title": "Scale Collaborates With NVIDIA to Bring Nemotron to Enterprises",
      "source_name": "scale-ai-blog",
      "source_url": "https://scale.com/blog/scale-partners-with-nvidia",
      "published_date": "2026-10-15"
    },
    {
      "manifest_id": "1792104492128212910",
      "title": "Post-Training Nemotron 3.5 Super VL for Long-Horizon Enterprise Work",
      "source_name": "prime-intellect-research",
      "source_url": "https://primeintellect.ai/blog/nemotron-3-5-super-vl",
      "published_date": "2026-10-15"
    },
    {
      "manifest_id": "1792099962120500714",
      "title": "Cline Agents Harness Unlocks Benchmark-Leading Performance from NVIDIA Nemotron 3.5 Super VL",
      "source_name": "cline-blog",
      "source_url": "https://cline.ghost.io/nvidia-nemotron-3-5-super-vl-cline-agent-harness",
      "published_date": "2026-10-15"
    },
    {
      "manifest_id": "1792113243663743217",
      "title": "One Year of Tinker: Turning Your Expertise into Custom Models",
      "source_name": "thinking-machines-lab-news",
      "source_url": "https://thinkingmachines.ai/news/one-year-of-tinker",
      "published_date": "2026-10-15"
    },
    {
      "manifest_id": "1792103185385117752",
      "title": "Deploy multimodal AI agents with Nemotron 3.5 Super VL",
      "source_name": "nscale-blog",
      "source_url": "https://nscale.com/blog/nvidia-nemotron-3-5-super-vl-available-on-nscale",
      "published_date": "2026-10-15"
    },
    {
      "manifest_id": "1792101863246132872",
      "title": "Announcing NVIDIA Nemotron 3.5 Super VL on Baseten",
      "source_name": "baseten-blog",
      "source_url": "https://baseten.co/blog/nvidia-nemotron-35-super-vl-launch",
      "published_date": "2026-10-15"
    }
  ],
  "run_these_manifests": {
    "title": "Run these Manifests yourself",
    "summary": "Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them.",
    "manifest_ids": [
      "1792097503049566042",
      "1792102448418904611",
      "1792104492128212910",
      "1792099962120500714",
      "1792113243663743217",
      "1792103185385117752",
      "1792101863246132872"
    ],
    "steps": [
      {
        "n": 1,
        "title": "Get a Synorb Key",
        "text": "Click Generate Free Agent Key at https://synorb.com/connect, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{\"client\":\"your-agent-name\"}`. The key is the `api_key` field of the answer. Save the key. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; 3 of them are starter Beacons."
      },
      {
        "n": 2,
        "title": "Read each cited Manifest by id",
        "text": "REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer <your Synorb Key>`. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief."
      },
      {
        "n": 3,
        "title": "Check the claim",
        "text": "Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them."
      }
    ],
    "mcp_url": "https://mcp.synorb.com/mcp",
    "connect_url": "https://synorb.com/connect",
    "request": {
      "method": "GET",
      "url_template": "https://api.synorb.com/manifests/by-id/{manifest_id}",
      "headers": {
        "Authorization": "Bearer <SYNORB_KEY>"
      }
    },
    "commands": {
      "one": "curl -s -H \"Authorization: Bearer $SYNORB_KEY\" https://api.synorb.com/manifests/by-id/1792097503049566042",
      "all": "for id in 1792097503049566042 1792102448418904611 1792104492128212910 1792099962120500714 1792113243663743217 1792103185385117752 1792101863246132872; do\n  curl -s -H \"Authorization: Bearer $SYNORB_KEY\" \"https://api.synorb.com/manifests/by-id/$id\"\ndone"
    }
  },
  "links": {
    "html": "https://hangingcontext.com/stories/2026-10-16/post-training-is-the-product/",
    "json": "https://hangingcontext.com/stories/2026-10-16/post-training-is-the-product/index.json",
    "markdown": "https://hangingcontext.com/stories/2026-10-16/post-training-is-the-product/index.md",
    "newsroom": "https://hangingcontext.com/n/artificial-intelligence/",
    "edition": "https://hangingcontext.com/editions/2026-10-16T12/"
  },
  "related": [
    {
      "url": "https://hangingcontext.com/stories/2026-10-15/agent-pricing-chasm/",
      "headline": "Everyone's deploying agents and nobody trusts them: the agentic chasm in five numbers",
      "slug": "agent-pricing-chasm"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-10-15/voice-ai-full-duplex-era/",
      "headline": "Voice AI's full-duplex era: xAI goes sub-second at $0.08 a minute",
      "slug": "voice-ai-full-duplex-era"
    },
    {
      "url": "https://hangingcontext.com/stories/2026-10-14/world-models-robotics/",
      "headline": "Robotics' new bet: models that predict the world instead of reacting to it",
      "slug": "world-models-robotics"
    }
  ],
  "sponsor": {
    "label": "From Synorb",
    "title": "Claims with a source link",
    "text": "A Manifest holds a Signal of structured claims and a Brief, plus the source it came from. Your agent can check a sentence against it.",
    "cta": {
      "label": "Read the guide",
      "url": "https://synorb.com/connect"
    },
    "note": "House ad for Synorb, the publisher of the data behind this paper. Not a news item."
  },
  "disclosure": "Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify.",
  "reuse": {
    "version": "1",
    "summary": "Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product.",
    "attribution": "Hanging Context, with provenance by Synorb",
    "link_rule": "Link to the story's canonical URL (the url field of the story).",
    "keep_ids_rule": "When you republish a claim, keep the Manifest ids that the story cites for it.",
    "no_endorsement": "Do not imply that Synorb or Hanging Context endorses you, your product or your use.",
    "license_url": "https://hangingcontext.com/terms/reuse/",
    "machine_url": "https://hangingcontext.com/.well-known/reuse.json",
    "canonical_url": "https://hangingcontext.com/stories/2026-10-16/post-training-is-the-product/"
  }
}