{
  "_meta": {
    "publication": "The Model Pulse",
    "schemaVersion": "2026.05.02",
    "generatedAt": "2026-09-08T22:45:10.960Z",
    "canonicalUrl": "https://brianletort.ai/industry/models/2026-W31",
    "markdownUrl": "https://brianletort.ai/industry/models/2026-W31/llm.md",
    "pdfUrl": "https://brianletort.ai/downloads/model-pulse-2026-W31.pdf",
    "treeUrl": "https://brianletort.ai/industry/tree",
    "sourceFile": "src/data/industry/models/2026-W31.ts"
  },
  "issue": {
    "slug": "2026-W31",
    "isoYear": 2026,
    "isoWeek": 31,
    "issueNumber": 15,
    "publishedAt": "2026-08-01",
    "cadence": "weekly",
    "periodLabel": "Week 31 of 2026",
    "bigRead": {
      "headline": "Open weights finally became deployable artifacts — and the closed stack answered with post-training and rate cards, not new flagships",
      "body": "Moonshot's Jul 27 Kimi K3 weight drop converts last week's promissory open-frontier watch item into a procurement fact: a 2.8T MoE with 104B active parameters, 896 experts selecting 16, hybrid 69 KDA + 24 Gated MLA layers, 1,048,576 context, and MoonViT-V2 — under a revenue-tiered Kimi K3 License that triggers a separate commercial deal for MaaS above $20M trailing-twelve-month revenue. Self-host evaluation can now start on artifacts; license diligence is the gate, not model quality.\n\nNo closed frontier model shipped. The closed stack competed on price and post-training instead. OpenAI cut GPT-5.6 Luna 80% to $0.20/$1.20 and Terra 20% to $2/$12 on Jul 30 while holding Sol at $5/$30 and replacing Priority Processing with Fast mode at 2x token price for up to ~2.5x speed. DeepSeek put V4-Flash-0731 into public beta on Jul 31 on the same 284B/13B architecture and unchanged $0.14/$0.28 pricing; Artificial Analysis independently scored it +10 Intelligence Index points to 50. Thinking Machines released Inkling-Small the same day as the Luna cut — 276B/12B active, Apache 2.0, independent AA Index 40 versus Inkling's 41 — with a context-window conflict to resolve (vendor card up to 1M; AA article lists 256K).\n\nThe week's largest vendor-claimed capability jump did not need new weights either: OpenAI reported ARC-AGI-3 public-set scores moving from 13.3% to 38.3% on GPT-5.6 Sol by retaining reasoning between tool calls and compacting context instead of truncating, with roughly 6x fewer output tokens. That is a harness result, not an official leaderboard reprint — and no qualifying LMArena, SWE-bench, or ARC official board moves landed in-window. Gemini 3.5 Pro missed its Jul 31 GA prediction. MiniMax H3 promised open weights and had not shipped them as of Aug 1. Net/net: route on measured Index and task cost, treat open as license-plus-weights, and re-test agents against memory policy before buying a bigger model."
    },
    "treeDelta": {
      "summary": "Inkling-Small joins the tree as the strongest US open-weights efficiency result to date, reaching an independent Intelligence Index of 40 against Inkling's 41 with under a third of the parameters. Kimi K3 is reclassified from closed to open weights on the July 27 artifacts and license text, with its active-parameter count corrected from a community estimate to a confirmed 104B. DeepSeek V4 Flash updates for the 0731 re-post-train checkpoint that gained 10 independent Index points on unchanged architecture.",
      "added": [
        "inkling-small"
      ],
      "updated": [
        "kimi-k3",
        "deepseek-v4-flash"
      ],
      "note": "Kimi K3 is an update rather than an addition: the row has existed since W29, and this week's material change is openness, license terms, and a corrected active-parameter count of 104B against the earlier community-estimated ~50B. Qwen3.7 Flash is held out of the tree pending a first-party model card, and MiniMax H3 is deferred until its promised weights actually ship. Both are covered in the cards."
    },
    "frontierMovements": [
      {
        "name": "GPT-5.6 Luna / Terra price cut and Sol Fast mode",
        "vendor": "OpenAI",
        "releaseDate": "2026-07-30",
        "headline": "Luna falls 80% to $0.20/$1.20, Terra 20% to $2/$12, Sol holds $5/$30, and Fast mode sells ~2.5x speed at 2x token price",
        "why": "Re-baseline high-volume routing immediately: Luna undercuts most Western cheap tiers on sticker price while Sol remains the premium band against Opus 5. Treat Fast mode as a purchasable latency tier rather than a separate model, and instrument per-tier completion rates before moving fleet volume on list price alone. The AI Stack Weekly owns the fixed-workload spread calculation; this Pulse stays on the architecture and routing implication.",
        "tier": "frontier",
        "architecture": "reasoning",
        "source": "OpenAI; Unite.AI; VentureBeat; The Decoder",
        "sourceUrl": "https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/"
      },
      {
        "name": "DeepSeek V4-Flash-0731",
        "vendor": "DeepSeek",
        "releaseDate": "2026-07-31",
        "headline": "Same 284B/13B MoE and $0.14/$0.28 pricing, re-post-trained into public beta; independent AA Index jumps 10 points to 50",
        "why": "Drop this into existing Flash routes as a cheap-tier agent upgrade without a size or price change — but do not treat it as V4-Pro GA, and pin the harness when comparing Terminal-Bench (vendor-claimed 82.7 max effort versus Artificial Analysis 79%). Architecture and model ID stay deepseek-v4-flash; Pro, app, and web remain unchanged pending the promised official V4-Pro.",
        "tier": "frontier",
        "architecture": "moe",
        "source": "DeepSeek API changelog; Artificial Analysis (independent)",
        "sourceUrl": "https://api-docs.deepseek.com/updates/"
      },
      {
        "name": "GPT-5.6 Sol harness memory on ARC-AGI-3",
        "vendor": "OpenAI",
        "releaseDate": "2026-07-29",
        "headline": "Vendor-reported public-set score triples from 13.3% to 38.3% via retained reasoning plus context compaction on the same model",
        "why": "Treat the ARC-AGI-3 jump as vendor-claimed on a single harness, not an official leaderboard reprint: OpenAI reports roughly 6x fewer output tokens when compaction replaces rolling truncation on the Responses API, and no qualifying LMArena or SWE-bench official moves landed in-window. Agent Techniques Weekly owns the method and the reproduction caveat.",
        "tier": "reasoning",
        "architecture": "agentic",
        "source": "OpenAI (vendor-reported harness result)",
        "sourceUrl": "https://openai.com/index/how-two-settings-tripled-our-arc-agi-3-scores/"
      },
      {
        "name": "Qwen3.7 Flash",
        "vendor": "Alibaba / Qwen",
        "releaseDate": "2026-07-27",
        "headline": "Quiet low-cost multimodal API at reported $0.03/$0.13 under 32K and 1M context — no lab blog, no weights, no public benchmark table",
        "why": "Useful as a fleet-cost floor candidate for multimodal agents, but treat specs and pricing as secondary until a first-party model card is pinned. Sources are OpenRouter listing dates and technical digests citing a QwenCloud changelog; no qwenlm.github.io launch post was found in-window.",
        "tier": "specialist",
        "architecture": "multimodal",
        "source": "eesel.ai; BenchLM (secondary; no primary lab announcement)",
        "sourceUrl": "https://benchlm.ai/models/qwen3-7-flash"
      }
    ],
    "openWeights": [
      {
        "modelId": "kimi-k3",
        "name": "Kimi K3",
        "vendor": "Moonshot AI",
        "releaseDate": "2026-07-27",
        "headline": "Full weights ship: 2.8T MoE, 104B active, 896/16 experts, hybrid 69 KDA + 24 Gated MLA, 1,048,576 context, revenue-tiered license",
        "why": "Start the self-host evaluation now — GitHub and Hugging Face artifacts landed Jul 27 — but send the Kimi K3 License to counsel before any commercial hosting plan: MaaS above $20M aggregate trailing-12-month revenue needs a separate deal, with attribution triggers above 100M MAU or $20M monthly revenue. Independent AA Index reference remains 57; vendor-claimed table includes GPQA Diamond 93.5 and DeepSWE 67.5. Hosted API stays $3/$15 (cached input $0.30).",
        "tier": "open_frontier",
        "architecture": "moe",
        "source": "MoonshotAI/Kimi-K3 GitHub; Hugging Face; license text; Unite.AI",
        "sourceUrl": "https://huggingface.co/moonshotai/Kimi-K3"
      },
      {
        "name": "Inkling-Small",
        "vendor": "Thinking Machines Lab",
        "releaseDate": "2026-07-30",
        "headline": "276B/12B-active multimodal MoE under Apache 2.0; independent AA Index 40 vs Inkling 41 at under a third the parameters",
        "why": "Strongest US open-weights efficiency play this week for teams that cannot absorb K3-class cluster scale: BF16 needs ~600 GB aggregate VRAM, NVFP4 ~180 GB, with vLLM v0.26.0 shipping same-week Inkling family support. Flag the context conflict honestly — vendor model card says up to 1M; Artificial Analysis lists 256K in its article blurb — and use Index 40 as the independent baseline rather than vendor-claimed SWE-bench Verified 80.2% alone.",
        "tier": "open_frontier",
        "architecture": "moe",
        "source": "Thinking Machines Lab; Artificial Analysis (independent Index 40); VentureBeat",
        "sourceUrl": "https://thinkingmachines.ai/news/inkling-small/"
      },
      {
        "name": "MiniMax H3",
        "vendor": "MiniMax",
        "releaseDate": "2026-07-31",
        "headline": "Multimodal video model launches API-first with open weights promised in coming days — still not on Hugging Face or GitHub as of Aug 1",
        "why": "Treat current state as API-only until the repo appears: unified text/image/video/audio understanding to video with native stereo, up to 15s at 2K, with vendor claims that 2K per-second price is under one-third of mainstream models. Do not build a self-host plan around the promise; the W30 K3 lesson applies again.",
        "tier": "specialist",
        "architecture": "multimodal",
        "source": "MiniMax blog; SCMP; secondary weights-not-yet status",
        "sourceUrl": "https://www.minimax.io/blog/minimax-h3"
      }
    ],
    "architectureWatch": [
      {
        "pattern": "Post-training and harness policy move scores without new weights",
        "examples": [
          "DeepSeek V4-Flash-0731: +10 AA Index on unchanged 284B/13B",
          "OpenAI ARC-AGI-3: 13.3% to 38.3% via retained reasoning + compaction",
          "OpenAI Jul 29 serving/harness efficiency post (~20% serving cost, >15% token-gen efficiency)"
        ],
        "body": "Two independent vendors moved measured capability this week without shipping new parameter counts: DeepSeek re-post-trained Flash in place, and OpenAI changed Responses API memory settings on the same Sol weights. Architects should treat reasoning retention, context compaction, and post-training checkpoints as first-class evaluation dimensions alongside the model ID. Buyers who only diff weight releases will miss the largest week-over-week gains.",
        "source": "DeepSeek; Artificial Analysis; OpenAI",
        "sourceUrl": "https://artificialanalysis.ai/articles/deepseek-v4-flash-0731-scores-50-on-the-artificial-analysis-intelligence-index-10-points-above-previous-deepseek-v4-flash"
      },
      {
        "pattern": "Revenue-tiered open licenses turn legal review into the deployment gate",
        "examples": [
          "Kimi K3 License: >$20M trailing-12-month MaaS commercial trigger",
          "Kimi K3 attribution triggers above 100M MAU or $20M monthly revenue",
          "Inkling-Small: clean Apache 2.0 contrast case"
        ],
        "body": "Downloadable weights no longer imply a single open-source diligence checklist. K3 is MIT-like until commercial hosting crosses the revenue trigger, at which point counsel must negotiate; Inkling-Small remains Apache 2.0 with no such trigger. Procurement should classify each open release by license economics before sizing clusters, and keep Apache-clean alternatives on the shortlist when MaaS revenue is material.",
        "source": "MoonshotAI/Kimi-K3 LICENSE; Thinking Machines Lab",
        "sourceUrl": "https://raw.githubusercontent.com/MoonshotAI/Kimi-K3/main/LICENSE"
      },
      {
        "pattern": "Serving stacks ship in the same week as the open MoE class",
        "examples": [
          "vLLM v0.26.0: Inkling modeling stack, NVFP4, MTP speculative decoding",
          "Inkling-Small BF16 ~600 GB / NVFP4 ~180 GB VRAM floors",
          "DeepSeek-V4 routing and speculative optimizations across NVIDIA/AMD/XPU"
        ],
        "body": "Production self-hosters got a same-week inference path for the new open MoE class rather than a multi-month serving lag. vLLM v0.26.0 (Jul 27) adds Inkling CUDA-graph and Hopper FA4 relative attention support alongside DeepSeek-V4 serving pushes and matured KV offloading. Capacity planners should price FP4 datapaths and KV-tiered storage into the model-release decision, not as a later optimization.",
        "source": "vLLM project GitHub release v0.26.0; Thinking Machines model card",
        "sourceUrl": "https://github.com/vllm-project/vllm/releases/tag/v0.26.0"
      },
      {
        "pattern": "Near-iso Intelligence at a fraction of active parameters",
        "examples": [
          "Inkling-Small: AA Index 40 vs Inkling 41 at <1/3 parameters",
          "Inkling-Small: 12B active vs Inkling 41B active",
          "DeepSeek V4 Flash prior: also sat at AA Index 40 in the same size class"
        ],
        "body": "Artificial Analysis places Inkling-Small within one Index point of full Inkling while activating 12B rather than 41B parameters, and notes no smaller open-weights model scores higher in that size class. Efficiency MoEs are now a procurement tier of their own: teams should ask whether the last Index point is worth 3x the serving footprint before defaulting to the flagship open checkpoint.",
        "source": "Artificial Analysis (independent)",
        "sourceUrl": "https://artificialanalysis.ai/articles/inkling-small-lands-within-a-point-of-inkling-on-the-artificial-analysis-intelligence-index-with-less-than-a-third-of-the-parameters"
      }
    ],
    "benchmarkMoves": [
      {
        "benchmark": "Artificial Analysis Intelligence Index — DeepSeek V4 Flash",
        "movement": "Independent Index jumps from 40 to 50 (+10) for Flash-0731 on unchanged architecture; now ~1 point behind GPT-5.6 Luna (51) and 6 points above V4 Pro on AA's comparison",
        "rows": [
          {
            "model": "DeepSeek V4 Flash 0731 (max)",
            "score": "50 Index; GDPval-AA v2 1559 Elo; Terminal-Bench 2.1 79% (AA)"
          },
          {
            "model": "Prior DeepSeek V4 Flash",
            "score": "40 Index; GDPval-AA v2 1189 Elo"
          },
          {
            "model": "Kimi K3 (max)",
            "score": "57 — open-frontier reference"
          },
          {
            "model": "GPT-5.6 Luna (max)",
            "score": "51"
          },
          {
            "model": "GLM-5.2 (max)",
            "score": "51"
          }
        ],
        "source": "Artificial Analysis (independent)",
        "sourceUrl": "https://artificialanalysis.ai/articles/deepseek-v4-flash-0731-scores-50-on-the-artificial-analysis-intelligence-index-10-points-above-previous-deepseek-v4-flash"
      },
      {
        "benchmark": "Artificial Analysis Intelligence Index — Inkling-Small",
        "movement": "New open US model lands at Index 40, within 1 point of Inkling (41) at less than a third the parameters",
        "rows": [
          {
            "model": "Inkling-Small",
            "score": "40 (independent)"
          },
          {
            "model": "Inkling",
            "score": "41 (independent)"
          },
          {
            "model": "DeepSeek V4 Flash (prior max, same size class)",
            "score": "40"
          },
          {
            "model": "MiniMax-M2.7 (per AA)",
            "score": "~38 — about 2 points below Inkling-Small"
          }
        ],
        "source": "Artificial Analysis (independent)",
        "sourceUrl": "https://artificialanalysis.ai/articles/inkling-small-lands-within-a-point-of-inkling-on-the-artificial-analysis-intelligence-index-with-less-than-a-third-of-the-parameters"
      },
      {
        "benchmark": "ARC-AGI-3 public set (vendor harness, not official board)",
        "movement": "OpenAI reports GPT-5.6 Sol rising from 13.3% to 38.3% by retaining reasoning and compacting context — same model, roughly 6x fewer output tokens",
        "rows": [
          {
            "model": "GPT-5.6 Sol + retained reasoning + compaction",
            "score": "38.3% (vendor-reported)"
          },
          {
            "model": "GPT-5.6 Sol + rolling truncation baseline",
            "score": "13.3% (vendor-reported)"
          }
        ],
        "source": "OpenAI (vendor-reported; not an official ARC leaderboard reprint)",
        "sourceUrl": "https://openai.com/index/how-two-settings-tripled-our-arc-agi-3-scores/"
      },
      {
        "benchmark": "DeepSeek Terminal-Bench 2.1 (vendor harness vs independent)",
        "movement": "Vendor-claimed max-effort 82.7 versus Artificial Analysis 79% on the same Flash-0731 checkpoint — pin the harness before ranking",
        "rows": [
          {
            "model": "DeepSeek V4 Flash 0731 (DeepSeek harness, max effort)",
            "score": "Terminal-Bench 2.1 82.7 (vendor-claimed)"
          },
          {
            "model": "DeepSeek V4 Flash 0731 (Artificial Analysis)",
            "score": "Terminal-Bench 2.1 79% (independent)"
          },
          {
            "model": "DeepSeek V4 Flash 0731 (DeepSeek harness)",
            "score": "Cybergym 76.7; Toolathlon verified 70.3; DeepSWE 54.4 (vendor-claimed)"
          }
        ],
        "source": "DeepSeek API changelog (vendor); Artificial Analysis (independent)",
        "sourceUrl": "https://api-docs.deepseek.com/updates/"
      }
    ],
    "scorecard": {
      "asOf": "2026-08-01",
      "rows": [
        {
          "tier": "Closed frontier",
          "leader": "Claude Fable 5",
          "challenger": "Claude Opus 5",
          "note": "No new closed flagship shipped; Fable retains the absolute reference while Opus 5 remains the economic default at $5/$25."
        },
        {
          "tier": "Open frontier",
          "leader": "Kimi K3",
          "challenger": "Inkling-Small",
          "note": "K3 takes the lead now that weights and license text are public (AA Index 57); Inkling-Small is the efficiency challenger at Index 40 under Apache 2.0."
        },
        {
          "tier": "Reasoning",
          "leader": "Claude Fable 5",
          "challenger": "GPT-5.6 Sol",
          "note": "Sol's vendor-reported ARC-AGI-3 harness jump shows memory policy can move scores without new weights; independent cost-per-task still pending."
        },
        {
          "tier": "Coding",
          "leader": "Claude Opus 5",
          "challenger": "DeepSeek V4 Flash 0731",
          "note": "Opus 5 holds the practical coding lead on W30 launch evidence; Flash-0731 is the cheap-tier agent upgrade at Index 50 and unchanged $0.14/$0.28."
        },
        {
          "tier": "Multimodal",
          "leader": "Gemini 3.6 Flash",
          "challenger": "Kimi K3",
          "note": "Gemini remains the public fleet multimodal default; K3 now adds self-host control with MoonViT-V2. Gemini 3.5 Pro still has not GA'd."
        },
        {
          "tier": "Edge / small",
          "leader": "Inkling-Small",
          "challenger": "Gemini 3.5 Flash-Lite",
          "note": "Inkling-Small brings open multimodal agents to ~180 GB NVFP4; Flash-Lite remains the hosted speed-cost floor at roughly 350 tok/s."
        }
      ]
    },
    "vendorSignals": [
      {
        "vendor": "OpenAI",
        "date": "2026-07-30",
        "signal": "GPT-5.6 Luna cut 80% to $0.20/$1.20 and Terra 20% to $2/$12; Sol held at $5/$30; Fast mode replaces Priority Processing",
        "meaning": "Closed labs are competing on completed-task economics via rate cards and latency tiers rather than new flagship weights this week. Route only latency-sensitive work into Fast mode and keep ordinary asynchronous agents on standard Luna or Terra after measuring completion quality.",
        "source": "OpenAI",
        "sourceUrl": "https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/"
      },
      {
        "vendor": "Moonshot AI",
        "date": "2026-07-27",
        "signal": "Kimi K3 License ships with weights: MIT-like until MaaS revenue exceeds $20M trailing twelve months",
        "meaning": "Open-weights diligence now splits into quality evaluation and license classification. Counsel should map expected MaaS and product revenue against the trigger before any production self-host commitment; Apache-clean alternatives remain available for teams that cannot accept the commercial clause.",
        "source": "MoonshotAI/Kimi-K3 LICENSE; Hugging Face",
        "sourceUrl": "https://raw.githubusercontent.com/MoonshotAI/Kimi-K3/main/LICENSE"
      },
      {
        "vendor": "Google DeepMind",
        "date": "2026-07-31",
        "signal": "Gemini 3.5 Pro still not generally available as the standing Jul 31 prediction deadline expires",
        "meaning": "Do not plan Q3 procurement around 3.5 Pro. Buy the callable Flash and Flash-Lite fleet now; another slip further weakens roadmap-based purchasing.",
        "source": "Google (Jul 21 partner-testing status; no in-window GA found)",
        "sourceUrl": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/"
      },
      {
        "vendor": "MiniMax",
        "date": "2026-07-31",
        "signal": "H3 launches with an explicit promise to open weights in the coming days; no public HF or GitHub artifacts as of Aug 1",
        "meaning": "Repeat the K3 lesson: an open-weights commitment is not a downloadable artifact. Keep H3 on the API watchlist until the repo and license appear.",
        "source": "MiniMax blog; secondary weights-not-yet reports",
        "sourceUrl": "https://www.minimax.io/blog/minimax-h3"
      },
      {
        "vendor": "vLLM project",
        "date": "2026-07-27",
        "signal": "vLLM v0.26.0 ships Inkling family support, NVFP4, MTP speculative decoding, and DeepSeek-V4 serving optimizations",
        "meaning": "Serving readiness is now part of the model-release race. Self-host teams evaluating Inkling-Small or DeepSeek V4 this week should pin v0.26.0 as the baseline inference path rather than waiting for a later stack catch-up.",
        "source": "vLLM GitHub releases",
        "sourceUrl": "https://github.com/vllm-project/vllm/releases/tag/v0.26.0"
      }
    ],
    "watchlist": [
      {
        "window": "Next 7-14 days",
        "title": "MiniMax H3 downloadable weights and license",
        "why": "Decides whether H3 becomes a real open video-generation option or remains an API product wrapped in open-language marketing."
      },
      {
        "window": "Next 30 days",
        "title": "DeepSeek V4-Pro official GA and pricing",
        "why": "Flash-0731 is not Pro GA; the standing surge-pricing and official Pro triggers remain incomplete until a callable Pro ID and rate card land."
      },
      {
        "window": "Next 14 days",
        "title": "Inkling-Small context window: vendor 1M vs AA 256K",
        "why": "Procurement specs need a resolved context number; watch for an AA methodology note or Thinking Machines clarification."
      },
      {
        "window": "By Aug 15",
        "title": "Independent Opus 5 cost-per-task results",
        "why": "Vendor benchmark leadership from W30 is still not enough; the market needs task cost, token use, and latency against Fable 5 and GPT-5.6 Sol."
      },
      {
        "window": "Next 2-3 weeks",
        "title": "LMArena Opus 5 Elo stabilization",
        "why": "No qualifying in-window official board move; secondary pages remain unstable — wait for a date-stamped stable print before re-ranking coding."
      },
      {
        "window": "Next 30 days",
        "title": "Gemini 3.5 Pro callable model ID and public pricing",
        "why": "The Jul 31 miss is now a track-record item; only a public GA closes the W30 prediction cleanly."
      }
    ],
    "changelog": [
      "W31 opens on the Kimi K3 weights-and-license conversion, Inkling-Small Apache 2.0 efficiency play, DeepSeek Flash-0731 independent +10 Index jump, and OpenAI Luna/Terra cuts plus harness ARC-AGI-3 result.",
      "Negative findings recorded as content: Gemini 3.5 Pro missed Jul 31 GA; no qualifying LMArena/SWE-bench/ARC official leaderboard moves; MiniMax H3 weights promised but not shipped.",
      "Tree delta lists kimi-k3 under added for this issue's openness conversion while models.yaml still awaits an Inkling-Small row; editor should restructure to added:[inkling-small] / updated:[kimi-k3] once the YAML lands."
    ]
  }
}