optimizing-for-cost-and-intelligence changedabout-claude/models/optimizing-for-cost-and-intelligence
Nearest release: v2.1.296, published under an hour after this site recorded the change. Shown because the two are within 24 hours of each other. Nothing here says the release caused the edit.
Recorded here
Lines+20added
Lines−20removed
From line
498
where the diff opens
First seen
14 Aug 2026
this site's first read of the page
Recorded edits18to this page, all time
The whole hunk
from line 498, old and new numbered
/
from line 498
498498 messages.append({"role": "user", "content": results})
499499```
500500
501On Claude Managed Agents, you can append the sentence to the task in the `user.message` event that starts the session, and the clock to each `user.message` and [`user.custom_tool_result`](https://platform.claude.com/docs/en/managed-agents/events-and-streaming#handling-custom-tool-calls) event you send. Turns that follow the platform's built-in tools, such as web search, see the last clock you sent, not the current time. In a [multiagent session](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration), the coordinator writes each worker's brief, so your client can't add the sentence to it. A worker sees a clock only in the results of custom tools that you run for it. Anthropic didn't measure the changes on Claude Managed Agents. To show the current time before every request, and to every agent in a team, run the agent loop yourself on the Messages API.
501On Claude Managed Agents, you can append the sentence to the task in the `user.message` event that starts the session, and the clock to each `user.message` and [`user.custom_tool_result`](https://platform.claude.com/docs/en/managed-agents/events-and-streaming#handling-custom-tool-calls) event you send. Turns that follow the platform's built-in tools, such as web search, see the last clock you sent, not the current time. In a [multiagent session](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration), the orchestrator writes each worker's brief, so your client can't add the sentence to it. A worker sees a clock only in the results of custom tools that you run for it. Anthropic didn't measure the changes on Claude Managed Agents. To show the current time before every request, and to every agent in a team, run the agent loop yourself on the Messages API.
502502
503503## Combine models
504504
from line 515
515515
516516In the advisor strategy, a lower-cost executor model runs the agent loop and performs most turns. When it hits a decision that needs deeper judgment, such as choosing an approach or recovering from a failure, it calls a higher-intelligence advisor model for strategic guidance, then continues. Most tokens are billed at executor rates, and only the occasional consultations at advisor rates.
517517
518To use it, add the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) to your request. This beta feature runs the whole strategy server-side in one `/v1/messages` request: the executor emits a tool call, Anthropic runs the advisor inference, and the executor continues with the advice; you write no orchestration code. On Claude Managed Agents, [give the session an advisor](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration#give-the-session-an-advisor) by adding an `advisor` entry to the agent's `multiagent` roster; the session's primary thread consults it the same way. Claude Code supports it too; see [escalating hard decisions with the advisor tool](https://code.claude.com/docs/en/advisor).
518To use it, add the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool) to your request. This beta feature runs the whole strategy server-side in one `/v1/messages` request: the executor emits a tool call, Anthropic runs the advisor inference, and the executor continues with the advice; you write no orchestration code. On Claude Managed Agents, [give the session an advisor](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration#give-the-session-an-advisor) by turning on the `advisor` setting in the agent's `multiagent` block; the session's primary thread consults it the same way. Claude Code supports it too; see [escalating hard decisions with the advisor tool](https://code.claude.com/docs/en/advisor).
519519
520520<Frame>
521521 
from line 553
553553
554554In the orchestrator strategy, the frontier model holds the loop. It decomposes the task, dispatches subtasks to lower-cost worker models, and merges their results. The orchestrator's own transcript stays short because workers absorb the token-heavy exploration, so most tokens are billed at worker rates while the plan and synthesis still come from the frontier model.
555555
556To build one, use [multiagent orchestration](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration) in Claude Managed Agents: configure a coordinator agent (the orchestrator) and a roster of worker agents, each with its own model. For a complete working example with a frontier coordinator and Claude Sonnet 5 workers, see the Claude Cookbook recipe [Coordinator pattern: big models for planning, small models for execution](https://github.com/anthropics/claude-cookbooks/blob/main/managed_agents/CMA_plan_big_execute_small.ipynb).
556To build one, use [multiagent orchestration](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration) in Claude Managed Agents: configure an orchestrator agent and worker agents that it can call, each with its own model.
557557
558558<Frame>
559559 
560560</Frame>
561561
562This pattern saves wall-clock time when workers can run in parallel: on the corpus benchmark[8](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), an episode took about 2.3 hours with the coordinator running the platform's documented limit of 25 concurrent workers, compared with 15 to 20 hours solo. It saved money in only two measured situations. On work a single model could handle alone, the same model at lower effort was cheaper every time.
562This pattern saves wall-clock time when workers can run in parallel: on the corpus benchmark[8](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), an episode took about 2.3 hours with the orchestrator running the platform's documented limit of 25 concurrent workers, compared with 15 to 20 hours solo. It saved money in only two measured situations. On work a single model could handle alone, the same model at lower effort was cheaper every time.
563563
564**Case 1: insurance against the cost tail on routine work.** A frontier model running alone occasionally spirals on a routine problem it would normally solve. Because you cannot tell in advance which those will be, a few such runs dominate the bill. A coordinator that hands routine work to a lower-cost worker caps that tail, because any spiraling now happens at worker rates.
564**Case 1: insurance against the cost tail on routine work.** A frontier model running alone occasionally spirals on a routine problem it would normally solve. Because you cannot tell in advance which those will be, a few such runs dominate the bill. An orchestrator that hands routine work to a lower-cost worker caps that tail, because any spiraling now happens at worker rates.
565565
566Anthropic measured this on a deliberately easy slice of BrowseComp[4](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) (10 problems the solo model reliably solves; 50 delegated and 70 solo runs). A Claude Fable 5 coordinator with one Claude Sonnet 5 worker cost about half as much as Claude Fable 5 alone on average and about a third as much at the 90th percentile ($12 USD compared with $33 USD), and the solo model's single most expensive run, at $84 USD, was also wrong:
566Anthropic measured this on a deliberately easy slice of BrowseComp[4](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) (10 problems the solo model reliably solves; 50 delegated and 70 solo runs). A Claude Fable 5 orchestrator with one Claude Sonnet 5 worker cost about half as much as Claude Fable 5 alone on average and about a third as much at the 90th percentile ($12 USD compared with $33 USD), and the solo model's single most expensive run, at $84 USD, was also wrong:
567567
568568<Frame>
569569 
from line 573
573573
574574**Case 2: work larger than one context window.** A solo model must work through an input that large serially, one context window at a time, paying to re-read its own state on every pass. Workers each read their own partition, in parallel and at worker rates. Reading-heavy work that still fits in one context window is a model-choice problem, not a delegation problem: on reading cost alone, the orchestrator comes out ahead only when no single context can hold the work.
575575
576Anthropic built a benchmark for this case[8](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs): a 21.6-million-token corpus of 14 public Python packages with 130 planted defects, too large for any context window. Lowering effort cannot help, because the bill is the corpus read itself: Claude Fable 5.1 solo cost $468 USD to $552 USD per episode across the three effort settings, and only its accuracy moved. The coordinator configuration, a Claude Fable 5.1 lead over 25 Claude Sonnet 5 workers, cost about half as much as those settings (47% to 55% less) and scored 10 to 12 points below them, in about 2.3 hours per episode against 15 to 20, while beating a Claude Sonnet 5 solo baseline outright:
576Anthropic built a benchmark for this case[8](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs): a 21.6-million-token corpus of 14 public Python packages with 130 planted defects, too large for any context window. Lowering effort cannot help, because the bill is the corpus read itself: Claude Fable 5.1 solo cost $468 USD to $552 USD per episode across the three effort settings, and only its accuracy moved. The orchestrator configuration, a Claude Fable 5.1 lead over 25 Claude Sonnet 5 workers, cost about half as much as those settings (47% to 55% less) and scored 10 to 12 points below them, in about 2.3 hours per episode against 15 to 20, while beating a Claude Sonnet 5 solo baseline outright:
577577
578578<Frame>
579 
579 
580580</Frame>
581581
582The token accounting shows the scale of the reading: the coordinator configuration read about 560 million cached tokens per episode, about one and a half times the solo model's roughly 365 million, nearly all of them at Claude Sonnet 5's cache-read rate, and still cost about half as much overall. Fable 5.1 at `high` effort still holds peak accuracy, at about 2.2 times the coordinator configuration's cost, so delegation here buys most of the accuracy, not all of it.
582The token accounting shows the scale of the reading: the orchestrator configuration read about 560 million cached tokens per episode, about one and a half times the solo model's roughly 365 million, nearly all of them at Claude Sonnet 5's cache-read rate, and still cost about half as much overall. Fable 5.1 at `high` effort still holds peak accuracy, at about 2.2 times the orchestrator configuration's cost, so delegation here buys most of the accuracy, not all of it.
583583
584**When delegation doesn't pay.** An orchestrator buys something only when there is bulk to hand off: many independent pieces, ideally too many for one context window. When the work is one dependent chain, or fits in a single context, the orchestrator pays for a plan, a handoff, and a merge that a single model gets for free. In every such case measured, the coordinator's model alone at lower effort came out ahead.
584**When delegation doesn't pay.** An orchestrator buys something only when there is bulk to hand off: many independent pieces, ideally too many for one context window. When the work is one dependent chain, or fits in a single context, the orchestrator pays for a plan, a handoff, and a merge that a single model gets for free. In every such case measured, the orchestrator's model alone at lower effort came out ahead.
585585
586The boundary is task difficulty, not the benchmark: on the full, harder BrowseComp[4](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) set, Claude Fable 5 alone reached the coordinator configuration's accuracy at 22% to 30% lower cost. Independent external work reports the same pattern[5](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs). If the work is one chain, fits in one context without a long cost tail, or a single model at lower effort already meets your bar, don't build an orchestrator.
586The boundary is task difficulty, not the benchmark: on the full, harder BrowseComp[4](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) set, Claude Fable 5 alone reached the orchestrator configuration's accuracy at 22% to 30% lower cost. Independent external work reports the same pattern[5](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs). If the work is one chain, fits in one context without a long cost tail, or a single model at lower effort already meets your bar, don't build an orchestrator.
587587
588588### Choose between the strategies
589589
from line 885
8858855. **Agent-architecture scaling:** Kim et al., "Towards a Science of Scaling Agent Systems," arXiv:2512.08296, 2025. Independent external study, cited only for the direction of the finding on when delegation does not pay, not for any figure.
8868866. **DeepWideSearch:** "DeepWideSearch: Benchmarking Depth and Width in Agentic Information Seeking," arXiv:2510.20168, 2025. The 220 questions span 15 domains, each combining many-row collection with multi-hop retrieval; measured on the benchmark's standing row set, 3 runs per configuration, run August 2, 2026 (the single-worker team point ran July 26 to 27, 2026).
8878877. **DeepResearch Bench II:** Li et al., "DeepResearch Bench II: Diagnosing Deep Research Agents via Rubrics from Expert Report," arXiv:2601.08536, 2026. Its 132 research tasks across 22 domains are graded against expert-derived binary rubrics; measured on a 50-task subset stratified across all themes, one attempt per task, 3 runs per setting, on [Claude Managed Agents](https://platform.claude.com/docs/en/managed-agents/overview) with the platform's own web search and fetch tools (August 26 to 27, 2026); scored on the 33 tasks no configuration refused, with attempts the production safety classifiers cut short removed; costs are what a customer is billed, the platform's requests plus web-search fees. Scores are each model's mean on the 33-task basis with its own pre-empted tasks removed; on the 21 tasks clean in every arm, Claude Fable 5.1 holds a 2-to-3-point lead over Claude Fable 5 at every effort level and both models are flat across effort. The caching chart re-prices the same requests with every input token at the uncached rate. Claude Opus 4.6 judges under the benchmark's rubric protocol; the original uses a different judge, and an Anthropic judge may favor the house style. Claude Opus 5 at its default effort ran on the same surface and subset, three runs, on August 28, 2026: 68.8% on the raw 50 tasks, 70.8% on the 33-task basis, and 71.1% on the 21-task set, at $6.71 USD per task ($23.72 USD without caching); none of its attempts was cut short by the safety classifiers, under a safeguards deployment newer than the one the other models ran under.
8888. **Corpus defect sweep:** Anthropic-internal, for work larger than one context window: a 21.6-million-token corpus from 14 public Python package sources with 130 planted defects and deterministic grading; protocol fixed before the runs and internally reviewed; three runs per configuration. Every configuration ran on Claude Managed Agents. The charted team configuration is a run in which the Claude Fable 5.1 coordinator ran the whole sweep inside the platform at its documented limit of 25 concurrent Claude Sonnet 5 workers, run August 30, 2026; its three episodes scored F1 0.764, 0.825, and 0.791 after the extras audit (raw 0.751, 0.821, and 0.781) for $225 USD, $234 USD, and $283 USD. The Claude Sonnet 5 solo configuration ran August 3 to 4, 2026; the Claude Fable 5.1 solo configurations ran August 24 to 25, 2026, under the platform's launch serving settings, three seeds per effort setting, on the same corpus build. The sandbox image carried installed copies of part of the corpus, and Claude Fable 5.1's final assembly step compared against them in 7 of 9 episodes; re-grading without those additions moved the affected seeds by up to 3 points. Absolute F1 is specific to this corpus build, not comparable across benchmarks; configuration comparisons are like for like.
8888. **Corpus defect sweep:** Anthropic-internal, for work larger than one context window: a 21.6-million-token corpus from 14 public Python package sources with 130 planted defects and deterministic grading; protocol fixed before the runs and internally reviewed; three runs per configuration. Every configuration ran on Claude Managed Agents. The charted team configuration is a run in which the Claude Fable 5.1 orchestrator ran the whole sweep inside the platform at its documented limit of 25 concurrent Claude Sonnet 5 workers, run August 30, 2026; its three episodes scored F1 0.764, 0.825, and 0.791 after the extras audit (raw 0.751, 0.821, and 0.781) for $225 USD, $234 USD, and $283 USD. The Claude Sonnet 5 solo configuration ran August 3 to 4, 2026; the Claude Fable 5.1 solo configurations ran August 24 to 25, 2026, under the platform's launch serving settings, three seeds per effort setting, on the same corpus build. The sandbox image carried installed copies of part of the corpus, and Claude Fable 5.1's final assembly step compared against them in 7 of 9 episodes; re-grading without those additions moved the affected seeds by up to 3 points. Absolute F1 is specific to this corpus build, not comparable across benchmarks; configuration comparisons are like for like.
8898899. **GPQA Diamond:** Rein et al., "GPQA: A Graduate-Level Google-Proof Q\&A Benchmark," 2023. The 198-question Diamond subset, two runs per configuration, run August 7, 2026 (Claude Opus 5.5: September 19, 2026), model-graded against reference answers, advisor tokens metered per request. A platform safety check refused two biology questions on the Claude Sonnet 5 executors, and one of them also on Claude Opus 5; excluding them changes no comparison by more than one point. Claude Opus 5.5's 92% comes from two runs that set `fallbacks: "default"` to opt into [server-side fallback](https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback#server-side-fallback), with any attempt that still ended in a refusal counted as wrong. In each run the safety check flagged six biology questions, Claude Opus 5 answered five of them through the fallback, and the sixth still ended in a refusal. Opus 5.5's cost per question includes those fallback answers. Without counting refusals as wrong, these runs score 93%, because the grader still assigns an answer option to a refused attempt, usually the correct one. With refusals counted as wrong, Claude Opus 5's runs score 91% (one refusal per run), as do two Claude Opus 5.5 runs with fallback off, in which Opus 5.5 refused five or six biology questions per run. Claude Haiku 5.5 and Claude Sonnet 5.5, each alone and with a Claude Opus 5.5 advisor, ran two runs per configuration on October 7, 2026, alongside two more Claude Opus 5.5 runs, all with fallback off, at each model's default effort, and with refusals counted as wrong. The safety check refused one biology question per run on Claude Haiku 5.5 and Claude Sonnet 5.5 alone, one in two runs on Haiku 5.5 with the advisor, none on Sonnet 5.5 with the advisor, and six per run on Opus 5.5, which scored 91% with them counted as wrong and 93.5% on the questions it answered. The advisor configurations added the tool without a prompt asking for consultations, and neither executor called the advisor on any question.
89089010. **DeepSWE:** Datacurve, "DeepSWE: Measuring Frontier Coding Agents on Original, Long-Horizon Engineering Tasks," arXiv:2607.07946, 2026. The set has 113 original tasks across five languages with program-based verifiers. Pairings are two runs each, run August 7, 2026, with advisor tokens metered per request, and used a client-side advisor loop rather than the advisor tool, with identical accounting. Single-model effort sweeps are single runs priced from token counts, a cache-aware approximation. Costs per task are run totals divided by 113.
89189111. **Internal agentic-coding benchmark:** Anthropic-internal: 370 repository tasks graded by the repositories' own tests. The API figures were measured with a 128,000-token output cap, one run per configuration: Opus 5 alone at the default effort August 9 to 10, 2026, and at `low` and `medium` August 10, 2026; Claude Fable 5.1 alone at five explicitly set effort values August 20, 2026 (the chart shows three of them); and the pairing August 24 to 25, 2026. Claude Opus 5.5 alone ran on all 370 tasks, September 19 to 20, 2026: at its default effort (`medium`) and at `high` with five attempts per task, and at `low` and `xhigh` with one (369 of 370 scored at each, after a setup-check failure). The Claude Opus 5.5 executor at `high` with the released Claude Fable 5.1 as advisor (the August runs used a pre-release snapshot) ran five attempts per task on the same dates; one task failed its setup check, so 1,845 attempts were scored. The advisor chart compares that pairing with the released Claude Fable 5.1 alone at `high`, its default, one attempt per task on October 7, 2026: 85.7%, 317 of 370 tasks. The 279 attempts in which the advisor was turned away under load were re-run, and attempts whose consults timed out were kept, as in August. The August runs had five attempts per task for the pairing and the Claude Opus 5 control and one for the other points. The August pairing averaged about two advisor consultations per attempt; the Claude Opus 5.5 pairing requested 1.39 and received 1.35. Costs are per attempt. Costs are priced as a customer's organization is metered: each agent-loop request's prior prompt as a cache read and its new tokens as a 5-minute cache write, from the runs' own usage records, and each advisor call, which uses no cache, from its recorded tokens, all at list prices. The Claude Code figures are runs of the same tasks from July 8 to 23, 2026, one run per configuration, costs approximate.
No line in this hunk matches that.