Claude workhorse for coding agents, careful analysis, and production cost control
Claude Sonnet 4.6 by anthropic is offered by 50 providers on this page. Public prices are shown for 47 of them. Its official list price is $3.00 per 1M input tokens and $15.00 per 1M output tokens. The lowest paid channel is Xpersona at $0.90 / $5.55 per 1M, about 4.0× below the list price. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. It also has 1 free ($0) channel; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.
Artificial Analysis rates it 30.1 on the Intelligence Index, with 63 for coding and 31.8 for agentic tasks. Against its lowest blended price of $2.06 per 1M, that is roughly 15 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 46 tokens per second, with 131.9s to the first token. Running one task of the Artificial Analysis suite costs about $2.4859, which reflects how many tokens its reasoning consumes rather than the unit price alone.
The context window is 1,000,000 tokens at the reference host, but hosts report different limits, from 200,000 to 1,000,000, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. Accepted input modalities are Text, Image, and PDF. Its training knowledge cuts off at 2025-08-31.
Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| Kenari | Gateway | Free | — | — | 1,000,000 | 64,000 | ||
| Xpersona | Gateway | $0.90 | $5.55 | $0.09 | — | 200,000 ⚠ | 128,000 | |
| Poe anthropic/claude-sonnet-4.6 | Gateway | $2.60 | $13.00 | $0.26 | $3.20 | 983,040 ⚠ | 128,000 | |
| Perplexity Agent | Gateway | $3.00 | $15.00 | $0.30 | — | 200,000 ⚠ | 64,000 | |
| NanoGPT anthropic/claude-sonnet-4.6 | Gateway | $3.00 | $15.00 | $0.30 | — | 1,000,000 | 128,000 | |
| AnthropicOfficial | First-party | $3.00 | $15.00 | $0.30 | $3.75 | 1,000,000 | 128,000 | |
| NEAR AI Cloud | Gateway | $3.00 | $15.00 | $0.30 | $3.75 | 1,000,000 | 64,000 | |
| Databricks databricks-claude-sonnet-4-6 | Cloud | $3.00 | $15.00 | $0.30 | $3.75 | 1,000,000 | 64,000 | |
| Kilo Gateway anthropic/claude-sonnet-4.6 | Gateway | $3.00 | $15.00 | $0.30 | $3.75 | 1,000,000 | 128,000 | |
| Abacus | Gateway | $3.00 | $15.00 | — | — | 1,000,000 | 64,000 | |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Interleaved thinking (reasoning between tool calls) is declared by 7 of 50 providers.
3 / 50 providers expose no reasoning control (reasoning_options: []).
1 more channels offer $0 (Kenari); free tiers usually have rate limits and no SLA, excluded from ranking.
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| SWE-Atlas Codebase QnA | harness: Claude Code | 31.2 | score | Source ↗ |
| SWE-Atlas Refactoring | harness: Claude Code | 32.21 | score | Source ↗ |
| SWE-Atlas Test Writing | harness: Claude Code | 31.76 | score | Source ↗ |
| Artificial Analysis Coding Agent Index | harness: Claude Code · variant: medium | 49.4 | average pass@1 | Source ↗ |
| SWE-Atlas Codebase QnA | harness: Claude Code · variant: medium | 70.3 | pass@1 | Source ↗ |
| SWE-Bench Pro | harness: Claude Code · variant: medium · dataset: hard-aa | 14.9 | pass@1 | Source ↗ |
| Terminal-Bench | harness: Claude Code · variant: medium · v2.1 | 63.1 | pass@1 | Source ↗ |
| Terminal-Bench | harness: Terminus-2 · v2.1 | 67 | success rate | Source ↗ |
| Humanity's Last Exam | variant: no tools | 34.6 | accuracy | Source ↗ |
| Humanity's Last Exam | variant: with tools | 46.8 | accuracy | Source ↗ |
| OSWorld-Verified | — | 78.5 | success rate | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.
Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and harness, so rows are not comparable with one another. The Intelligence Index above draws on a different, newer set of evaluations.