Provider
Anthropic
Base model family
Model intelligence profile
Claude Sonnet 4.6 benchmark results and model details
This family profile brings together 10 published benchmark records for Anthropic Claude Sonnet 4.6. Every result keeps its original benchmark, configuration, scale, and source; unrelated scores are never averaged.
Published benchmark evidence available · Business Skills V3 not yet authorized
Release date
2026-02-17
Exact-family source only
Published benchmarks
10
Native scales kept separate
Artificial Analysis
Reference available
As of 2026-07-16
Benchmark-level evidence
Published Claude Sonnet 4.6 benchmark results
These are individual benchmarks, not collection rollups. Scores remain on their original scales, and agent or harness results stay labelled as system configurations.
Spring Prompt benchmark
ROASBench
21.21
Average ROASBench score
A 12-month performance-marketing simulation scored on business outcomes, planning, behavior, and persona fit.
1 published configuration
Anthropic: Claude Sonnet 4.6
21.21
- Average contribution profit
- $-146,915
- Average ROAS
- 1.18×
- Completed runs
- 3
- Score variability
- ±1.57
Source: Spring Prompt
Recorded 25 Jul 2026
Published ROASBench cache · 12-month simulation
Source record →
source-native benchmark
Artificial Analysis Intelligence Index
34*
A composite index of language-model performance across agentic work, coding, scientific reasoning, knowledge, and long-context reasoning.
1 published configuration
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
34*
- Status
- Estimated by source
- Recorded
- 30 Jul 2026
Source: Artificial Analysis
Recorded 30 Jul 2026
Intelligence Index v4.1 · 2026-07-30
Source record →
source-native benchmark
Coding Agent Configurations
35.2
Coding Agent Index v1.2 point score
Artificial Analysis Coding Agent Index v1.2 point scores and task-specific operational measurements for exact agent and model configurations.
These are exact agent-plus-model configurations, not bare-model scores.
1 published configuration
Claude Code - Sonnet 4.6 (medium)
35.2
- Mean cost per task
- $1.97
- Mean wall time
- 13.7 min
source-native benchmark
GPQA Diamond
80%
The most challenging subset of Graduate-Level Google-Proof Q&A, focused on scientific reasoning.
1 published configuration
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
80%
- Status
- Measured
- Recorded
- 30 Jul 2026
Source: Artificial Analysis
Recorded 30 Jul 2026
Independent evaluation · Public leaderboard snapshot · 2026-07-30
Source record →
source-native benchmark
Humanity's Last Exam
11%
A broad expert-level benchmark of difficult academic reasoning and knowledge questions.
1 published configuration
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
11%
- Status
- Measured
- Recorded
- 30 Jul 2026
Source: Artificial Analysis
Recorded 30 Jul 2026
Independent evaluation · Public leaderboard snapshot · 2026-07-30
Source record →
source-native benchmark
IFBench
42%
An instruction-following benchmark with diverse, verifiable out-of-domain output constraints.
1 published configuration
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
42%
- Status
- Measured
- Recorded
- 30 Jul 2026
Source: Artificial Analysis
Recorded 30 Jul 2026
Independent evaluation · Public leaderboard snapshot · 2026-07-30
Source record →
source-native benchmark
MMMU-Pro
69%
A multimodal academic reasoning benchmark designed to reduce shortcuts and guessing across many disciplines.
1 published configuration
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
69%
- Status
- Measured
- Recorded
- 30 Jul 2026
Source: Artificial Analysis
Recorded 30 Jul 2026
Independent evaluation · Public leaderboard snapshot · 2026-07-30
Source record →
source-native benchmark
Repository Issue Resolution
74.4%
OpenHands SWE-Bench resolved
Direct OpenHands SWE-Bench outcomes for resolving real repository issues with a pinned OpenHands agent and language-model configuration.
Scores are official OpenHands SWE-Bench resolved percentages for exact OpenHands and model configurations, not bare-model scores or SpringPrompt predictions. Twenty-nine exact task sidecars support marginal 95% bootstrap score intervals; five rows explicitly have no interval. No full-cohort rank confidence is claimed.
1 published configuration
OpenHands v1.11.5 + claude-sonnet-4-6
74.4%
- 95% task-bootstrap interval
- 70.6–78.2%
- Task evidence
- exact-headline-reconstruction
Source: OpenHands Index
Recorded 30 Jun 2026
OpenHands · SWE-Bench 2026.06.30-3015ac6
Source record →
source-native benchmark
Structured Output Reliability
85.44%
Direct benchmark score
Directly measured ability to return accurate values in the requested structured schema across the benchmark's evaluated text, image and audio modalities.
Point order reproduces the source's direct Overall score. Rank ranges come from overlap of marginal record-cluster bootstrap intervals for Overall; they are not simultaneous confidence intervals for rank.
1 published configuration
Claude-Sonnet-4.6
85.44%
- 95% interval
- 84.94–85.93
- Evaluated modality coverage
- 100%
Source: Structured Output Benchmark (SOB)
Recorded 17 Jul 2026
Sob-v1@da785a8521c8954283b2989d01e54d80c4e023c6:upstream-provider-configs:temperature-0-where-supported:max-output-2048:reasoning-disabled-or-minimum-where-required:official-modality-weights
Source record →
source-native benchmark
τ²-Bench Telecom
79%
A legacy agentic tool-use benchmark for conversational work in a telecom environment.
1 published configuration
Claude Sonnet 4.6 (Non-reasoning, Low Effort)
79%
- Status
- Measured
- Recorded
- 30 Jul 2026
Source: Artificial Analysis
Recorded 30 Jul 2026
Independent evaluation · Public leaderboard snapshot · 2026-07-30
Source record →
Independent operational reference
Artificial Analysis details
Representative configuration: Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort). Lifecycle: deprecated; released 2026-02-17.
Input price
$3.00 / 1M tokens
Output price
$15.00 / 1M tokens
Median output speed
46.9 tok/s
Operational reference only. Price and throughput are not model-quality scores and never affect Spring Prompt comparisons.
Early-user signal
Initial community opinions
A paraphrased editorial synthesis of 4 Reddit discussions created in the 14 days after launch (17 Feb 2026 to 3 Mar 2026). Anecdotal context only — it never affects scores or rankings.
The search-indexed launch-window sample highlighted Claude Sonnet 4.6's speed, price, task focus, and occasional strong coding output. Recurring complaints involved superficial reasoning, distraction, instruction failures, hallucinations, and a less conversational style. Reports varied enough that early usefulness appeared closely tied to task size and the surrounding product.
Independent early-test signal
Early technical field tests
An editorial paraphrase of 2 launch-window field tests from 2 independent authors on X, including 2 reports with a described method or inspectable artifact. The window runs from 17 Feb 2026 up to 3 Mar 2026, and exact model identity was editor reviewed.
Early Sonnet 4.6 tests covered conditional routing, scheduling, contract workflows, and repeated Claude Code use. Workflow automation looked capable, but a practical coding comparison found that longer reasoning and more tool calls could slow end-to-end completion despite respectable token speed. The launch evidence supports broad utility with a real throughput tradeoff.
These reports are selectively surfaced and are not a representative sample. They never affect benchmark scores, rankings, winners, or comparison outcomes.
Business Skills V3 · proposed
How Spring Prompt plans to test Claude Sonnet 4.6
The setup below is proposed and may change until preflight and execution approval are complete. Existing benchmark evidence above does not authorize or stand in for a Business Skills V3 result.
- Proposed reasoning
- High reasoning
- Provider revision
- Exact provider revision will be resolved and frozen only after preflight and execution approval
- Tools and service
- No tools · Anthropic standard routing
- Planning configuration
- claude-sonnet-4.6-high
- Generation controls
- Provider-managed reasoning; no temperature override proposed
Future first-party coverage
15 proposed Business Skills V3 task areas
These links describe evaluation contracts, not published Claude Sonnet 4.6 results.
Professional Knowledge Work & StrategyBusiness Workflow AutomationCustomer Support & Service ResolutionSales & MarketingLegal Document WorkFinance, Accounting & SpreadsheetsData Analysis & SQLResearch & Evidence SynthesisBusiness Writing & EmailDocument Comprehension & ExtractionSummarization & Grounded KnowledgeSoftware Engineering & CodingFrontend, UI & Web AppsTool Calling & Structured OutputPresentations, Charts & Visual Communication
Model comparisons
Compare Claude Sonnet 4.6 side by side
Editorially reviewed comparisons appear first. Every page matches only benchmark records with the same reviewed protocol key and does not manufacture an overall winner.
Publication safeguards
What must pass before a V3 result appears
- 1First-party task-local comparisons and eligible external evidence must both be present.
- 2Model and provider configuration identity must match the reviewed release exactly.
- 3Coverage, reliability, judge diagnostics, and sealed stability checks must pass.
- 4Uncertainty and missing evidence remain visible when results are published.
Stable family URL
Evidence can grow without changing the page
New reviewed benchmark snapshots, operational facts, community themes, and early field-test syntheses can be added here while the canonical model-family identity remains fixed.
More from Anthropic