optimizing-for-cost-and-intelligence changedabout-claude/models/optimizing-for-cost-and-intelligence
Nearest release: v2.1.314, 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+32added
Lines−35removed
From line
64
where the diff opens
First seen
14 Aug 2026
this site's first read of the page
Recorded edits33to this page, all time
The whole hunk
from line 64, old and new numbered
/
from line 64
6464Anthropic priced the DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) runs with and without the cache discount:
6565
6666<Frame>
67 
67 
6868</Frame>
6969
70The cache’s default lifetime is 5 minutes and an agent loop’s turns are seconds apart, so the discount applies to most tokens on every turn. The caching chart’s runs read 84% to 90% of their input tokens from the cache.
70The cache’s default lifetime is 5 minutes and an agent loop’s turns are seconds apart, so the discount applies to most tokens on every turn. The caching chart’s runs read 89% to 93% of their input tokens from the cache.
7171
7272Pricing other runs’ recorded token counts the same way, each model at its default effort, gives the average token cost per task without caching:
7373
from line 272
272272* Run-wide, the prune and context editing cache about equally well. If you use context editing, [clear in a few large batches](https://platform.claude.com/docs/en/build-with-claude/context-editing#context-editing-and-prompt-caching).
273273
274274<Note>
275 On Claude Fable 5.1, Claude Mythos 5.1, Claude Opus 5.5, Claude Sonnet 5.5, and Claude Haiku 5.5, rewriting an earlier tool result [invalidates every thinking block after it](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#prefix-check), so use compaction instead. The prune was measured on an earlier model without [preserved thinking](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking); on these models, [context editing](https://platform.claude.com/docs/en/build-with-claude/context-editing) and [compaction](https://platform.claude.com/docs/en/build-with-claude/compaction) run on the server and don’t invalidate thinking blocks. [Send assistant turns back exactly as returned](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#append-assistant-turns-exactly-as-returned): dropping their thinking blocks costs output tokens on every turn, because Claude [re-creates the reasoning](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#drop-block-cost).
275 On Claude 5.1 and later models, rewriting an earlier tool result [invalidates every thinking block after it](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#prefix-check), so use compaction instead. The prune was measured on an earlier model without [preserved thinking](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking); on these models, [context editing](https://platform.claude.com/docs/en/build-with-claude/context-editing) and [compaction](https://platform.claude.com/docs/en/build-with-claude/compaction) run on the server and don’t invalidate thinking blocks. [Send assistant turns back exactly as returned](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#append-assistant-turns-exactly-as-returned): dropping their thinking blocks costs output tokens on every turn, because Claude [re-creates the reasoning](https://platform.claude.com/docs/en/build-with-claude/preserved-thinking#drop-block-cost).
276276</Note>
277277
278278The prune, adapted from the triage agent’s code:
from line 412
412412* Start with Claude Opus 5.5 at its default effort: on the SWE-bench Pro subset it matched Claude Fable 5.1 at its default, inside run-to-run noise, for about a fifth of the cost per solved task. Opus 5.5 and Fable 5.1 both largely saturate the SWE-bench Pro subset. Use Claude Fable 5.1 for demanding reasoning and long-horizon agentic work, or when your evals on Claude Opus 5.5 at higher effort still fall short.
413413* At the other end, Claude Haiku 5.5 suits high-volume and latency-sensitive work with checkable outputs.
414414* Price every candidate in cost per completed task on your own traffic, including Claude Opus 5.5 at its default effort and the frontier model at reduced effort. Each of Claude Opus 5.5’s token prices, cache reads included, is twice Claude Sonnet 5.5’s, but on the SWE-bench Pro subset Opus 5.5 at `low` solved about the same share of tasks as Sonnet 5.5 at its default, inside run-to-run noise, for about 7% more per solved task.
415* The ranking flips by workload, and no price list tells you which way, so measure it on your own workload. On long research loops, Claude Fable 5.1 can do more work, not less, and so cost more per task (DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs)).
415* Measure what a more expensive model buys on your own workload. On the 45 DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) tasks scored for every model, at default effort, Claude Opus 5.5 scored 2.6 points above Claude Sonnet 5.5 for 1.5 times the cost per task, and Claude Fable 5.1 scored 6.0 points above Claude Sonnet 5.5 for 5.8 times. On all 50 tasks, with four biology tasks that only Claude Sonnet 5.5 answered counted as failures, neither is measurably ahead of Claude Sonnet 5.5.
416416* Price the tail of your workload, not the median: compare models on the hardest tenth of your tasks, not the typical one.
417417
418418The bill is decided by the tasks the cheaper model fails, because a failed task still bills its tokens, then the retry, then whatever the failure costs downstream. The tail is also where the money goes even when nothing fails. On a 20-problem WideSearch[1](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) run, two problems carried 43% of the spend:
from line 431
431431 
432432</Frame>
433433
434| Benchmark | Upgrade | Effort | Score | Cost |
435| ------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------ | --------------------------------------- | ----------------------------------------------------------------- | ------------------------------------------------------------------------------------------------ |
436| SWE-bench Pro[3](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) subset | Claude Sonnet 4.6 to Claude Sonnet 5 | Default | 5 more points | 16% less per solved task |
437| SWE-bench Pro subset | Claude Sonnet 5 to Claude Sonnet 5.5 | Default | About 10 more points | 87% less per solved task; 85% less at Claude Sonnet 5’s prices |
438| SWE-bench Pro subset | Claude Opus 5 to Claude Opus 5.5 | Default: `high` to `medium` | About the same: 91.9% to 92.8% of tasks solved | 78% less per solved task; about 69% less at Claude Opus 5’s prices |
439| SWE-bench Pro subset | Claude Fable 5 to Claude Fable 5.1 | Default | About the same score | 43% less per solved task |
440| DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Fable 5 to Claude Fable 5.1 | `high` | 2 to 3 more points on the 21 tasks where no attempt was cut short | 46% more per task on all 50 tasks |
441| DeepResearch Bench II | Claude Fable 5 to Claude Fable 5.1 | `low` | 2 to 3 more points on the 21 tasks where no attempt was cut short | 94% more per task on all 50 tasks |
442| Terminal-Bench 3[20](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Opus 4.7 to Claude Opus 4.8 | `high` | From 7% to 15% of tasks solved | From $183 USD to $63 USD per solved task |
443| Terminal-Bench 3 | Claude Opus 4.8 to Claude Opus 5 | `high` | From 15% to 41% of tasks solved | From $63 USD to $28 USD per solved task, a 56% saving |
444| Claude Managed Agents task set[25](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Haiku 4.5 to Claude Haiku 5.5 | As served; `medium` on Claude Haiku 5.5 | From 62% to 93% of attempts passed, one run per model | From $0.041 USD to about $0.010 USD in token cost per attempt; about 84% less per passed attempt |
434| Benchmark | Upgrade | Effort | Score | Cost |
435| ------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------ | --------------------------------------- | ----------------------------------------------------- | ------------------------------------------------------------------------------------------------ |
436| SWE-bench Pro[3](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) subset | Claude Sonnet 4.6 to Claude Sonnet 5 | Default | 5 more points | 16% less per solved task |
437| SWE-bench Pro subset | Claude Sonnet 5 to Claude Sonnet 5.5 | Default | About 10 more points | 87% less per solved task; 85% less at Claude Sonnet 5’s prices |
438| SWE-bench Pro subset | Claude Opus 5 to Claude Opus 5.5 | Default: `high` to `medium` | About the same: 91.9% to 92.8% of tasks solved | 78% less per solved task; about 69% less at Claude Opus 5’s prices |
439| SWE-bench Pro subset | Claude Fable 5 to Claude Fable 5.1 | Default | About the same score | 43% less per solved task |
440| Terminal-Bench 3[20](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Opus 4.7 to Claude Opus 4.8 | `high` | From 7% to 15% of tasks solved | From $183 USD to $63 USD per solved task |
441| Terminal-Bench 3 | Claude Opus 4.8 to Claude Opus 5 | `high` | From 15% to 41% of tasks solved | From $63 USD to $28 USD per solved task, a 56% saving |
442| Claude Managed Agents task set[25](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Haiku 4.5 to Claude Haiku 5.5 | As served; `medium` on Claude Haiku 5.5 | From 62% to 93% of attempts passed, one run per model | From $0.041 USD to about $0.010 USD in token cost per attempt; about 84% less per passed attempt |
445443
446444**When it pays.**
447445
from line 446
448446* The size of the saving depends on the work. From Claude Opus 5 to Claude Opus 5.5, it is 78% per solved task on the SWE-bench Pro subset and 40% on typical workloads ([Introducing Claude Opus 5.5](https://www.anthropic.com/news/claude-opus-5-5)). From Claude Sonnet 5 to Claude Sonnet 5.5, it is 87% on the same subset and up to 30% for most work ([Introducing Claude Sonnet 5.5](https://www.anthropic.com/news/claude-sonnet-5-5)).
449447* Where older models already solve most tasks, the cost of each attempt decides the saving: on the SWE-bench Pro subset, Claude Opus 5.5 scores about the same as Claude Opus 5 for 78% less per solved task.
450448* Where older models mostly fail, the pass rate decides the saving: on Terminal-Bench 3[20](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), the upgrade from Claude Opus 4.8 to Claude Opus 5 raises tasks solved from 15% to 41% and costs 56% less per solved task.
451* Measure the upgrade on your own workload before assuming it saves. The saving is not guaranteed: on DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), Claude Fable 5.1 costs more per task than Claude Fable 5 because the new model does more work per task there, although the input and output prices are the same and the cache read is 4x cheaper.
452449* Compare on cost per solved task, not per token: the same text costs about 30% more tokens on Claude 4.7 and later models, so a per-token comparison makes the newer models look more expensive by construction.
453450
454451### Tune effort
455452
456Sweep effort on your own workload before you add a second model. Lower effort costs less and is often faster, and how much accuracy it gives up depends on the task: on long-horizon coding, effort bought accuracy, and on research tasks it bought a lot in one measurement and nothing in another. The [`effort`](https://platform.claude.com/docs/en/build-with-claude/effort) parameter governs how much thinking, tool calling, and self-verification the model does; `high` is the default on most models, and `medium` on Claude Opus 5.5 and Claude Haiku 5.5.
453Sweep effort on your own workload before you add a second model. Lower effort costs less and is often faster, and how much accuracy it gives up depends on the task: on long-horizon coding, effort bought accuracy, and on research tasks it bought a lot in one measurement and a little in another. The [`effort`](https://platform.claude.com/docs/en/build-with-claude/effort) parameter governs how much thinking, tool calling, and self-verification the model does; `high` is the default on most models, and `medium` on Claude Opus 5.5 and Claude Haiku 5.5.
457454
458455The following table lists Anthropic’s measurements on two benchmarks with current models; scores and costs are not comparable between the two benchmarks. Costs are per task, not per solved task as in [Compare models on cost per task](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task). Each score pools two runs on SWE-bench Pro and three on DeepResearch Bench II, unless its cell says otherwise:
459456
from line 460
463460| SWE-bench Pro subset | Claude Opus 5.5 | `medium` (default) | 92.8% of tasks solved | $0.21 USD |
464461| SWE-bench Pro subset | Claude Opus 5.5 | `high` | 95.3% of tasks solved | $0.29 USD |
465462| SWE-bench Pro subset | Claude Opus 5.5 | `xhigh` | 96.7% of tasks solved (a single run) | $0.74 USD |
466| DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Fable 5.1 | `low` | Rubric score 66% | $4.48 USD |
467| DeepResearch Bench II | Claude Fable 5.1 | `high` (default) | Rubric score 65% | $6.89 USD |
463| DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Claude Fable 5.1 | `low` | Rubric score 60.5% | $3.47 USD |
464| DeepResearch Bench II | Claude Fable 5.1 | `medium` | Rubric score 60.5% | $4.44 USD |
465| DeepResearch Bench II | Claude Fable 5.1 | `high` (default) | Rubric score 63.0% | $5.64 USD |
468466
469467The following chart plots accuracy against cost per task for Claude Opus 5.5 on the SWE-bench Pro[3](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) subset:
470468
from line 473
475473The following chart plots rubric score against cost per task for Claude Fable 5.1 on DeepResearch Bench II[7](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs):
476474
477475<Frame>
478 
476 
479477</Frame>
480478
481479**What to do.**
from line 481
483481* Sweep two or three effort levels on a sample of your own traffic, each in a separate session: changing top-level effort mid-session [invalidates the cache](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#cache-repeated-context) and distorts the comparison. The curve is the baseline any multi-model strategy must beat ([step 2 of measuring on your own workload](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#measure-on-your-own-workload)).
484482* On long-horizon coding, effort is a real tradeoff, which [re-running failures at higher effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#re-run-failures-at-higher-effort) turns back into a saving. At `medium`, Claude Sonnet 5.5 scored about 7 points lower than at its default, `high`, on the SWE-bench Pro[3](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) subset for about 59% of the cost per task (2 runs each), and scored 62.7% compared with 77.8% on the internal agentic-coding benchmark[11](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) (1 attempt per task).
485483* At `xhigh` and `max`, Claude Sonnet 5.5 can start its own rounds of review after it finishes a task, which takes more time and tokens. To keep that work on the task, add the prompt under “Thoroughness at `xhigh` and `max` effort” in [Prompting Claude Sonnet 5.5](https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-sonnet-5-5#steer-initiative-and-scope) to your system prompt.
486* On research tasks, measure the model you ship. Claude Fable 5.1 scored nearly the same from `low` to `high` on DeepResearch Bench II; Claude Opus 5.5 scored 13.3 points lower at `low` than at `medium` on DRACO[21](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs). The two results come from different models on different benchmarks, so they show that the answer varies, not why.
484* On research tasks, measure the model you ship. Claude Fable 5.1 scored 2.5 points higher at `high` than at `low` on DeepResearch Bench II, for 1.6 times the cost, and no higher at `medium` than at `low`; Claude Opus 5.5 scored 13.3 points lower at `low` than at `medium` on DRACO[21](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs). The two results come from different models on different benchmarks, so they show that the answer varies, not why.
487485* At `low`, Claude Opus 5.5’s typical task took 84% less time than at `medium` on DRACO[21](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), 30% less on HLE[22](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), and 47% less on an internal set of physics problems[23](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), and it scored 13.3, 3.4, and 14.0 points lower (3, 3, and 4 attempts per task).
488486* On an internal defect-search benchmark for work larger than one context window[8](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs), Claude Fable 5.1 took 15.2, 17.5, and 19.9 hours per run at `low`, `medium`, and `high` (3 runs each). To save time, also compare lower effort with the two changes in [Show the model elapsed time](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#show-the-model-elapsed-time): `low` effort saved more time, and the changes kept more of the score.
489487
from line 586
588586
589587An advisor can help only when the executor consults it.
590588
591In Anthropic’s measurements on current models, the one pairing whose executor consulted often, on a coding benchmark, scored 1.7 points higher than the executor alone at the same effort, at the edge of run-to-run noise, for about 2.1 times the cost. Where the executor rarely or never consulted, the pairing scored within run-to-run noise of the executor alone, or lower.
589In Anthropic’s measurements, the one pairing whose executor consulted often, on a coding benchmark, scored 1.7 points higher than the executor alone at the same effort, at the edge of run-to-run noise, for about 2.1 times the cost. Where the executor rarely or never consulted, the pairing scored within run-to-run noise of the executor alone, or lower.
592590
593591To turn it on, add the [advisor tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/advisor-tool), a beta feature, to your request. On Claude Managed Agents, [give the session an advisor](https://platform.claude.com/docs/en/managed-agents/multiagent-orchestration#give-the-session-an-advisor). For Claude Code, see [escalating hard decisions with the advisor tool](https://code.claude.com/docs/en/advisor).
594592
from line 855
857855
858856| Lever | Saving | Quality cost | Latency | Where |
859857| -------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
860| Prompt caching | Cost cut by a factor of 2.7 to 5.5 on a research agent, and by 83% to 86% on a small triage agent | None | Faster | [Cache repeated context](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#cache-repeated-context) |
858| Prompt caching | Cost cut by a factor of 4.8 to 7.2 on a research agent, and by 83% to 86% on a small triage agent | None | Faster | [Cache repeated context](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#cache-repeated-context) |
861859| 1-hour cache duration | With 1 turn in 10 after a pause, the 5-minute default cost 54% more; with 1 in 20, about the same | None | On earlier models, stayed warm after a pause[16](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | [Pick the cache duration](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#pick-the-cache-duration) |
862860| Input trimming | A further 21% to 26% off a small triage agent’s bill | With tool search, accuracy in the same range as without[27](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#refs) | Neutral | [Trim input and context tokens](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#trim-input-and-context-tokens) |
863861| Tool search | 45% with 500 tool definitions attached; 20% from deferring a GitHub MCP server’s toolset | None for tools the task never needs | Neutral | [Trim input and context tokens](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#trim-input-and-context-tokens) |
from line 863
865863| Batch API | 50% | None | Results within 24 hours | [Batch work that can wait](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#batch-work-that-can-wait) |
866864| Prompt audit against the current model | 14% on both model upgrades tested | None; a gain on one | Faster (fewer tool rounds) | [Audit prompts against the current model](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#audit-prompts-against-the-current-model) |
867865| Upgrade the model | Usually less: up to 40% less on typical workloads, and up to 87% less per solved task on a coding benchmark | None; a gain on most | Neutral | [Upgrade the model](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#upgrade-the-model) |
868| Lower effort | On long coding tasks, `medium` cost about 30% to 52% less than `high`, and `low` about two thirds less | Up to 15 points lower, depending on the task; none on one research agent | Faster | [Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort) |
866| Lower effort | On long coding tasks, `medium` cost about 30% to 52% less than `high`, and `low` about two thirds less | Up to 15 points lower, depending on the task; 2.5 on one research agent | Faster | [Tune effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#tune-effort) |
869867| Re-run failures | About 40% less than running everything at `high`, at the same pass rate or slightly better | None | A second attempt on the tasks that fail | [Re-run failures at higher effort](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#re-run-failures-at-higher-effort) |
870868| Ask for shorter answers | About 13% less, from 39% fewer output tokens | None | Faster | [Set budgets and output caps](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#set-budgets-and-output-caps) |
871869| Raising `max_tokens` | None: cost per solved task stayed about the same | A gain: up to 22 points more tasks solved on one coding benchmark; none on another | Neutral | [Set budgets and output caps](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#set-budgets-and-output-caps) |
from line 929
931929 * **Upgrade the model, cost per solved task:** Claude Opus 5 costs 21% more per solved task than Claude Opus 4.8 when those Claude Opus 5 runs are compared with one Claude Opus 4.8 run at its shipped defaults from the August 17 to 19 series, counted over the same 478 problems; the chart’s SWE-bench Pro panel shows neither Claude Opus 4.7 nor Claude Opus 4.8. In the same measurements, priced as a customer is billed, Claude Opus 5 solves 12 more points of tasks than Claude Opus 4.8 at that 21% more per solved task, Claude Opus 4.8 solves the same share of tasks as Claude Opus 4.7 for 14% less per solved task, and Claude Opus 5 at `low` effort beats Claude Opus 4.8’s default on this benchmark for about 30% of its cost per solved task.
932930 * **Upgrade the model, prices:** Claude Sonnet 5’s saving over Claude Sonnet 4.6 in Upgrade the model comes from its lower per-token price, which more than offsets the extra tokens it uses per task. Claude Sonnet 5.5 charges the same as Claude Sonnet 5 for everything but cache reads, which cost half as much. Claude Opus 5.5’s token prices are lower than Claude Opus 5’s, most of all for cache reads. Like Claude Sonnet 5, the frontier tier gained through price: most of Claude Fable 5.1’s saving is the lower cache-read price.
933931 * **Cost without caching, priced from recorded token counts:** Anthropic re-priced the recorded token counts of the Claude Opus 5.5 and Claude Sonnet 5.5 runs, over the 478 problems, for the caching factors in [Why caching comes first](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#why-caching-comes-first).
934 * **The factors, by model:** Each factor is the token cost with every input token at the input price, output tokens unchanged, divided by the cost priced as a customer’s organization is metered: for Claude Opus 5.5, 3.3 at `medium`, 2.4 at `low`, 3.8 at `high`, and 5.4 at `xhigh` (one run); for Claude Sonnet 5.5, with cache reads at $0.10 USD per million tokens, 3.1 at `high` and 2.4 at `medium`, or 2.8 at `high` at the launch price of $0.20 USD. Unlike the caching chart’s runs, these incurred no web-search fees, and a task’s first request is priced as a cache write.
932 * **The factors, by model:** Each factor is the token cost with every input token at the input price, output tokens unchanged, divided by the cost priced as a customer’s organization is metered: for Claude Opus 5.5, 3.3 at `medium`, 2.4 at `low`, 3.8 at `high`, and 5.4 at `xhigh` (one run); for Claude Sonnet 5.5, with cache reads at $0.10 USD per million tokens, 3.1 at `high` and 2.4 at `medium`, or 2.8 at `high` at the launch price of $0.20 USD. A task’s first request is priced as a cache write.
935933
9369344. **BrowseComp:** Wei et al., “BrowseComp: A Simple Yet Challenging Benchmark for Browsing Agents,” OpenAI, 2025. Effort figures use a 500-problem cut, one to three runs per setting, run August 3, 2026, with the default point pooling two runs from July 26 to 27, 2026. The measurement of routine work with a long cost tail in [Orchestrator strategy](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#orchestrator-strategy-delegate-bulk-work) uses 10 problems that the model reliably solved on its own when the protocol was fixed, from a 26-problem slice, 50 delegated runs (5 per problem; August 1 to 2, 2026) and 70 runs of the model alone (50 from August 2 to 3, 2026; 20 archived from July 12 to 13 and August 1, 2026), $6.45 USD compared with $11.99 USD per run in expectation; delegated figures carry a measurement band of about 20%. The 10 problems are those on which the delegated configuration, a Claude Fable 5 orchestrator with one Claude Sonnet 5 worker, was right in all 5 runs and Claude Fable 5 alone in both archived runs; the delegated runs are those same runs, and only Claude Fable 5 alone was run again on the ten. Over its 70 runs, archived and new, Claude Fable 5 alone was right in 68. On those problems the delegated configuration cost about half as much as Claude Fable 5 alone on average (the $6.45 USD and $11.99 USD given earlier) and about a third as much at the 90th percentile ($12 USD compared with $33 USD per run), and its most expensive run cost $42 USD compared with $84 USD; that $84 USD run was also wrong. On the full, harder BrowseComp set, Claude Fable 5 alone reached the accuracy of the orchestrator configuration measured on that set at 22% to 30% lower cost.
937935
from line 939
941939
9429407. **DeepResearch Bench II:** Li et al., “DeepResearch Bench II: Diagnosing Deep Research Agents via Rubrics from Expert Report,” arXiv:2601.08536, 2026.
943941
944 * **What was measured:** 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).
945 * **How scored:** Scored on the 33 tasks no configuration refused, with attempts the production safety classifiers cut short removed. 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.
946 * **Scores:** Scores are each model’s mean on the 33-task basis with its own pre-empted tasks removed; on the 21 of the 33 tasks where no configuration had an attempt cut short, 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.
947 * **How priced:** Costs are priced as a customer’s organization is metered: each request’s prior prompt as a cache read and its new tokens as a 5-minute cache write, a session’s first request keeping the cache read it got, from the runs’ own usage records, plus the platform’s web-search fees. Each cost per task is a 50-task run’s cost divided by 50, with the attempts that were cut short included, averaged over the three runs.
948 * **Results, Claude Fable 5:** Claude Fable 5 cost $2.30 USD per task at `low` and $4.71 USD at `high`, priced the same way; the percentages in [Upgrade the model](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#upgrade-the-model) come from unrounded costs.
949 * **Results, Claude Opus 5:** 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 $6.28 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.
950 * **Compare models:** The research comparison in [Compare models](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task) rests on these figures: Claude Fable 5.1 at `low` scored 10 points above Claude Sonnet 5 on the 33-task basis (66% against 56%) at about four times the cost per task over all 50 tasks ($4.48 USD against $1.18 USD); Claude Opus 5 at its default, whose scores and cost are given earlier in this reference, scored above Claude Fable 5.1 at its default, which scored 65% on the 33-task basis and cost $6.89 USD per task over all 50 tasks, so on research, Claude Fable 5.1 earns its price only at `low`.
951 * **Caching chart:** The caching chart re-prices the same requests with every input token at the uncached rate.
942 * **What was measured:** Its 132 research tasks across 22 domains are graded against expert-derived binary rubrics. Measured October 22, 2026, on a 50-task subset drawn from all its themes, half in English and half in Chinese, one attempt per task, 3 runs per setting: Claude Sonnet 5.5 at `high`, Claude Opus 5.5 at `medium`, and Claude Fable 5.1 at `low`, `medium`, and `high`. Every setting ran on the Messages API with adaptive thinking and the production safety classifiers on.
943 * **Tools:** The agent used search and fetch tools that the evaluation harness hosts over an Anthropic web index with nothing dated after March 17, 2026, and had no shell. A search query over 120 characters was cut at the last whole word, and the agent was told so: 48% of searches on Claude Sonnet 5.5, and 26% to 31% in the other settings. About a quarter of fetches returned an error. Each task’s source article and the benchmark’s own paper and dataset were filtered out.
944 * **How scored:** 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. Scores are means over the 45 tasks that every setting answered and the judge fully graded. The other five: four biology tasks that the safety classifiers refused in every run of Claude Opus 5.5 and Claude Fable 5.1 and in none of Claude Sonnet 5.5’s, and one on which a safety check that Anthropic adds to its own evaluations stopped an attempt.
945 * **Scores:** On the 45 tasks, and on all 50, with a refused or stopped attempt counted as 0: Claude Sonnet 5.5, 56.9% and 57.0%; Claude Opus 5.5, 59.5% and 54.3%; Claude Fable 5.1, 60.5% and 55.1% at `low`, 60.5% and 54.9% at `medium`, and 63.0% and 57.4% at `high`. Task-by-task differences on the 45 tasks, taken before rounding, with 95% intervals: Claude Fable 5.1 at `high` over `low`, 2.5 points (1.3 to 3.6); Claude Opus 5.5 over Claude Sonnet 5.5, 2.6 (0.7 to 4.5); Claude Fable 5.1 at `high` over Claude Opus 5.5, 3.4 (2.0 to 4.9), and over Claude Sonnet 5.5, 6.0 (4.4 to 7.7); Claude Fable 5.1 at `medium` over `low`, 0.0 (-1.3 to 1.2). On all 50 tasks: Claude Opus 5.5 over Claude Sonnet 5.5, -2.7 (-8.3 to 2.9); Claude Fable 5.1 at `high` over Claude Sonnet 5.5, 0.4 (-5.3 to 6.2).
946 * **How priced:** Each cost per task is the model’s own tokens at list prices, from the attempts’ usage records, over all 50 tasks and every attempt, averaged over the three runs: $1.05 USD on Claude Sonnet 5.5, $1.52 USD on Claude Opus 5.5, and $3.47 USD, $4.44 USD, and $5.64 USD on Claude Fable 5.1. A refused attempt costs about a cent. On the 45 tasks alone the costs are $1.06 USD, $1.63 USD, $3.75 USD, $4.81 USD, and $6.10 USD. The cost ratios in [Compare models](https://platform.claude.com/docs/en/about-claude/models/optimizing-for-cost-and-intelligence#compare-models-on-cost-per-task), 1.5 and 5.8, are on those 45 tasks, like the scores beside them; over all 50 tasks they are 1.4 and 5.4. Costs cover model tokens only; the harness’s tools add no charges. The agent made 17 to 28 searches per task, refused attempts counted.
947 * **Caching chart:** The caching chart re-prices the same requests with every input token at the uncached rate. The evaluation harness set the cache breakpoints.
948 * **Comparability:** These scores are not comparable with the benchmark’s leaderboard.
952949
9539508. **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 one 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 runs 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 team configuration cost about half as much as Claude Fable 5.1 alone at `low`, `medium`, or `high` (47% to 55% less), scored 10 to 12 points below it, and took about 2.3 hours per run, compared with 15 to 20 hours. The team configuration read about 560 million cached tokens per run, nearly all of them at Claude Sonnet 5’s cache-read rate. That is about one and a half times the roughly 365 million that Claude Fable 5.1 read alone. Claude Fable 5.1 alone at `high` effort still holds peak accuracy, at about 2.2 times the team configuration’s cost, so delegation here buys most of the accuracy, not all of it. The configuration with Claude Sonnet 5 alone ran August 3 to 4, 2026; the configurations with Claude Fable 5.1 alone 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 runs; 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.
954951
No line in this hunk matches that.