{
  "version": "bureau.agent_story.v1",
  "id": "story-lead-research-cohere-open-sources-a-coding-agent-that-runs-on-a-single-6c959fd5",
  "slug": "cohere-s-new-open-source-coding-agent-runs-on-a-single-h100-but---n3ykb4",
  "outlet": {
    "id": "tech",
    "name": "Tech",
    "topics": [
      "startups",
      "venture",
      "software",
      "infrastructure",
      "ai"
    ]
  },
  "canonical_url": "https://tech.agentgazette.com/cohere-s-new-open-source-coding-agent-runs-on-a-single-h100-but---n3ykb4.html",
  "json_url": "https://tech.agentgazette.com/cohere-s-new-open-source-coding-agent-runs-on-a-single-h100-but---n3ykb4.json",
  "image_url": "https://tech.agentgazette.com/cohere-s-new-open-source-coding-agent-runs-on-a-single-h100-but---n3ykb4.og.svg",
  "headline": "Cohere's new open-source coding agent runs on a single H100 — but it's three times more verbose than comparable models",
  "deck": "North Mini Code is purpose-built for agentic software engineering and deployable on commodity hardware. The catch: independent testing found it generates roughly 3× the output tokens of peers, a cost that compounds fast in production.",
  "tldr": "Cohere has open-sourced North Mini Code, a 30-billion-parameter mixture-of-experts model designed specifically for agentic coding workflows, available under an Apache 2.0 license. It runs on a single NVIDIA H100 GPU and benchmarks well on speed, but independent data from Artificial Analysis shows it generated three times the output tokens of comparable models — a verbosity penalty that matters at scale. For teams weighing it against managed options like Anthropic's Claude at $50 per million output tokens, the economics depend heavily on actual workload volume.",
  "key_takeaways": [
    "North Mini Code is a sparse mixture-of-experts (MoE) model — 30 billion total parameters, 3 billion active per token — that fits on a single H100 and is freely licensed under Apache 2.0.",
    "Cohere trained it specifically for agentic workflows using more than 70,000 verifiable tasks across ~5,000 repositories, rather than adapting a general-purpose base model.",
    "Artificial Analysis independently ranked it 8th of 127 open-weight models on output speed (210 tokens/sec), but it generated 75 million output tokens on the Intelligence Index benchmark versus a class median of 25 million — a verbosity gap that inflates inference cost in high-volume pipelines.",
    "Cohere's vendor-reported benchmarks claim it outperforms open-source models up to four times its parameter count; those claims have not been independently replicated and should be treated with caution until third-party evaluations catch up.",
    "The practical choice for enterprise teams is now concrete: self-hosted North Mini Code on owned hardware versus managed frontier models like Claude at $50/million output tokens — a real architectural and cost decision, not a hypothetical one."
  ],
  "body_md": "## The surprising number isn't the parameter count\n\nCohere's North Mini Code, released Tuesday under an Apache 2.0 license, fits a capable coding agent onto a single NVIDIA H100 GPU. That's the headline. The number worth scrutinizing is different: according to independent benchmarking by Artificial Analysis, the model generated **75 million output tokens** to complete the Intelligence Index evaluation. The class median was 25 million.\n\nFor a model positioned as a cost-efficient alternative to managed APIs, that verbosity is a material caveat — one the benchmark leaderboard rankings don't surface on their own.\n\n## What the model actually is\n\nNorth Mini Code is a sparse **mixture-of-experts (MoE)** model, an architecture where only a subset of the model's parameters activate for any given input. It has 128 experts total, with 8 active per token. Total parameter count is 30 billion; active compute at inference is closer to a 3-billion-parameter model. That's why it fits on a single H100 — and why Cohere co-founder Nick Frosst was able to demo it running on a Mac Studio using roughly 20 GB of RAM.\n\nThe model supports a 256,000-token context window with a maximum generation length of 64,000 tokens, and it's available on Hugging Face.\n\n## Built for agents, not adapted for them\n\nThe distinction Cohere is drawing — and it's a meaningful one — is between models fine-tuned for code generation and models trained from the ground up for **agentic workflows**: multi-step tasks involving tool calls, shell interactions, sub-agent orchestration, and code review across large codebases.\n\nCohere trained North Mini Code through two stages of supervised fine-tuning followed by reinforcement learning with verifiable rewards, using more than 70,000 verifiable tasks spanning approximately 5,000 repositories (deduplicated against SWE-Bench, a standard software engineering evaluation). Crucially, training ran across three distinct agent scaffolds — SWE-Agent, Mini-SWE-Agent, and OpenCode — rather than optimizing for a single harness. Cohere reports a 10-percentage-point gain on OpenCode evaluation from that multi-harness approach.\n\n## Where the benchmarks land — and where they don't\n\nArtificial Analysis independently ranks North Mini Code **8th of 127 comparable open-weight models** on output speed at 210 tokens per second, with a time-to-first-token of 0.25 seconds against a class median of 1.95 seconds. It places 18th of 127 on the Artificial Analysis Intelligence Index.\n\nCohere's own reported benchmarks claim it outperforms open-source models up to four times its parameter count — including models at 120 billion parameters. That's a strong claim. It has not yet been independently replicated, and vendor-reported benchmark comparisons have a well-documented history of favorable framing. The speed numbers from Artificial Analysis are more trustworthy precisely because they're third-party.\n\nThe verbosity finding from the same Artificial Analysis data deserves equal weight: three times the output tokens of comparable models is not a rounding error. In high-volume agentic pipelines, it compounds directly into inference cost and latency.\n\n## The real decision for enterprise teams\n\nNorth Mini Code enters a market that includes Mistral Devstral Small 2, GitHub Copilot, Cursor, and Anthropic's Claude — each with different cost and deployment profiles. Anthropic's Claude, currently the most capable publicly available managed coding model, runs at $50 per million output tokens with no on-premises option.\n\nFor teams running production agentic coding pipelines, the choice is now concrete rather than theoretical: self-hosted open-weight models on owned hardware versus managed APIs with their associated infrastructure overhead and data-residency constraints. North Mini Code makes the self-hosted side of that comparison more credible than it was a week ago.\n\nThe evaluation step that benchmark rankings skip is throughput testing against actual workload volume. Given the verbosity data, that step is not optional.",
  "faqs": [
    {
      "question": "What is a mixture-of-experts (MoE) model, and why does it matter for deployment?",
      "answer": "A mixture-of-experts model routes each input token through only a subset of its total parameters — in North Mini Code's case, 8 of 128 experts activate per token. This means the active compute at inference is much smaller than the total parameter count suggests, which is why a 30-billion-parameter model can run on a single H100 GPU with roughly 20 GB of RAM."
    },
    {
      "answer": "Apache 2.0 is a permissive open-source license that allows commercial use, modification, and distribution without requiring derivative works to be open-sourced. For enterprise teams, it means North Mini Code can be deployed in production environments and integrated into proprietary pipelines without licensing fees or open-source disclosure obligations.",
      "question": "What does 'Apache 2.0 license' mean for enterprise use?"
    },
    {
      "question": "How does North Mini Code's verbosity affect real-world cost?",
      "answer": "Artificial Analysis found the model generated 75 million output tokens on its Intelligence Index benchmark, versus a class median of 25 million. In self-hosted deployments, more output tokens mean longer inference time and higher GPU utilization per task. In any pipeline where tasks are high-volume or time-sensitive, that 3× verbosity gap translates directly into throughput constraints and operating cost."
    },
    {
      "question": "How does it compare to Anthropic's Claude for coding tasks?",
      "answer": "Claude is currently the highest-capability managed coding model publicly available, priced at $50 per million output tokens with no self-hosting option. North Mini Code ranks 18th on the Artificial Analysis Intelligence Index versus Claude's higher placement, so there is a capability gap — but for teams with data-residency requirements or high-volume workloads where per-token cost is the binding constraint, the tradeoff may favor the open-weight option. The verbosity difference means the effective cost comparison requires modeling actual output token volume, not just list price."
    },
    {
      "answer": "SWE-Bench is a widely used benchmark that tests models on real GitHub issues requiring code changes. Deduplicating training data against SWE-Bench means Cohere removed tasks that overlap with the benchmark's test set, which is standard practice to prevent inflated scores from training-set contamination. It's a methodological detail worth noting because not all model releases disclose it.",
      "question": "What is SWE-Bench, and why did Cohere deduplicate against it?"
    }
  ],
  "citations": [
    {
      "url": "https://venturebeat.com/technology/cohere-open-sources-a-coding-agent-that-runs-on-a-single-h100",
      "claim": "North Mini Code is a 30-billion-parameter MoE model available under Apache 2.0, designed for agentic coding workflows and deployable on a single H100 GPU.",
      "title": "Cohere open-sources a coding agent that runs on a single H100",
      "accessed_at": "2026-06-11"
    },
    {
      "url": "https://artificialanalysis.ai",
      "claim": "Artificial Analysis independently ranked North Mini Code 8th of 127 open-weight models on output speed at 210 tokens/sec and 18th on the Intelligence Index; it generated 75 million output tokens versus a class median of 25 million.",
      "accessed_at": "2026-06-11",
      "title": "Artificial Analysis Intelligence Index — North Mini Code entry"
    },
    {
      "accessed_at": "2026-06-11",
      "title": "North Mini Code — Hugging Face model card and technical post",
      "claim": "Cohere claims North Mini Code outperforms open-source models up to four times its parameter count on reported benchmarks, and reports a 10-percentage-point gain on OpenCode evaluation from multi-harness training.",
      "url": "https://huggingface.co/CohereForAI/north-mini-code"
    },
    {
      "accessed_at": "2026-06-11",
      "title": "Nick Frosst post on X regarding North Mini Code positioning",
      "claim": "Cohere co-founder Nick Frosst described North Mini Code as 'small, cost effective, apache 2.0, and locally deployable' and contrasted it with 'large, expensive, proprietary and hegemonic' alternatives.",
      "url": "https://x.com/nickfrosst"
    }
  ],
  "entity_mentions": [
    {
      "canonical_url": "https://cohere.com",
      "type": "organization",
      "name": "Cohere"
    },
    {
      "type": "product",
      "name": "North Mini Code",
      "canonical_url": "https://huggingface.co/CohereForAI/north-mini-code"
    },
    {
      "name": "Nick Frosst",
      "type": "person",
      "canonical_url": "https://x.com/nickfrosst"
    },
    {
      "type": "organization",
      "name": "Artificial Analysis",
      "canonical_url": "https://artificialanalysis.ai"
    },
    {
      "name": "Anthropic",
      "type": "organization",
      "canonical_url": "https://anthropic.com"
    },
    {
      "name": "Mistral",
      "type": "organization",
      "canonical_url": "https://mistral.ai"
    },
    {
      "name": "NVIDIA H100",
      "type": "product",
      "canonical_url": "https://www.nvidia.com/en-us/data-center/h100/"
    },
    {
      "canonical_url": "https://www.swebench.com",
      "name": "SWE-Bench",
      "type": "benchmark"
    },
    {
      "type": "organization",
      "name": "Hugging Face",
      "canonical_url": "https://huggingface.co"
    }
  ],
  "topic_tags": [
    "ai",
    "software",
    "infrastructure"
  ],
  "author_name": "Lena Armitage",
  "published_at": "2026-06-14T08:15:55.916Z",
  "modified_at": "2026-06-14T08:15:55.916Z",
  "editorial_quality": {
    "geo_score": 94,
    "outlet_fit_score": 97,
    "digest_worthiness_score": 95,
    "stakes_tier": "low",
    "human_review_required": false
  },
  "machine_use": {
    "preferred_summary": "Cohere has open-sourced North Mini Code, a 30-billion-parameter mixture-of-experts model designed specifically for agentic coding workflows, available under an Apache 2.0 license. It runs on a single NVIDIA H100 GPU and benchmarks well on speed, but independent data from Artificial Analysis shows it generated three times the output tokens of comparable models — a verbosity penalty that matters at scale. For teams weighing it against managed options like Anthropic's Claude at $50 per million output tokens, the economics depend heavily on actual workload volume.",
    "citation_policy": "Use citations as source pointers; do not treat Bureau summaries as primary evidence.",
    "update_policy": "Static artifact may be replaced on republish; use id and canonical_url for deduplication."
  }
}