The most transparent reasoning model on the market — ideal when you need to see and trust the thought process, not just the answer.
93
Coding
62
Writing
85
Research
0
Images
38
Value
82
Long Context
Use this when
Tackling complex coding challenges, mathematical proofs, and multi-step logical problems where visible reasoning and higher accuracy matter more than speed.
Skip this if
You need fast, high-volume responses or are doing straightforward writing, summarization, or conversational tasks where extended thinking adds cost and latency without meaningful quality gains.
Pricing
$3.00/1M in
$15.00/1M out
→0%since May 2026
Context
200k tokens
Speed
Deliberate
Thinking tokens (the internal reasoning trace) count toward output token billing, which can significantly increase costs on complex queries. The thinking budget can often be configured via the API. Best used selectively for tasks that genuinely benefit from deliberation rather than as a default model.
Visible chain-of-thought reasoning that can be inspected and debugged, unlike black-box reasoning models
Exceptional performance on hard coding tasks — competitive with o3-mini on SWE-bench style benchmarks
200K context window supports large codebases, long documents, and multi-turn agentic sessions
More reliable on logic-heavy tasks than standard Claude 3.7 Sonnet without the thinking overhead of o1-pro or Claude 3 Opus
Weaknesses
Thinking tokens add latency and cost — extended reasoning makes it noticeably slower than standard Sonnet and much slower than flash-tier models
At $15/1M output tokens, thinking traces can inflate bills significantly on high-volume or long-context tasks
Not suited for creative writing or casual chat where the deliberation overhead adds no value
Real-world use cases
What people actually use Claude 3.7 Sonnet (thinking) for.
Debugging a complex async race condition in a TypeScript microservices architecture with full reasoning trace
Working through a multi-constraint optimization problem in operations research with step-by-step verification
Analyzing a 150-page legal contract for risk clauses with structured logical reasoning across the full document
How Claude 3.7 Sonnet (thinking) compares
The nearest models people weigh against it, and what actually separates them.
vs Claude 3.5 Sonnet — Against Claude 3.5 Sonnet (Anthropic), Claude 3.7 Sonnet (thinking) runs about 50% cheaper per token and answers slower. Which one wins depends on whether context depth or latency is your constraint.
vs Claude Fable 5 — Against Claude Fable 5 (Anthropic), Claude 3.7 Sonnet (thinking) runs about 70% cheaper per token and gives up 5x on context. Take Claude 3.7 Sonnet (thinking) unless you specifically need what Claude Fable 5 does better.
vs Claude Fable 5.1 — Against Claude Fable 5.1 (Anthropic), Claude 3.7 Sonnet (thinking) runs about 70% cheaper per token and gives up 5x on context. Take Claude 3.7 Sonnet (thinking) unless you specifically need what Claude Fable 5.1 does better.
Price History
Claude 3.7 Sonnet (thinking) pricing over time
→0% since May 30
90 data points · tracked daily since May 30, 2026
Ready to try it?
Start using Claude 3.7 Sonnet (thinking)
Tackling complex coding challenges, mathematical proofs, and multi-step logical problems where visible reasoning and higher accuracy matter more than speed.. Start free — no card required.
Claude 3.5 Sonnet is Anthropic's mid-cycle flagship model, balancing strong reasoning, coding, and instruction-following with a 200K context window. It sits between Haiku and Opus in Anthropic's lineup, offering near-flagship quality at a lower cost than top-tier models.
Verdict
One of the best models for coding and complex instruction-following, but its premium pricing demands premium use cases.
Quality score
81%
Pricing
$6.00/1M in
$30.00/1M out
Speed
Balanced
3/5 speed
Context
200k tokens
Pricing at $6 input / $30 output per million tokens is significantly higher than GPT-4o ($2.50/$10). Best accessed via Anthropic API or Amazon Bedrock. Claude 3.5 Sonnet (October 2024 version) supersedes the June 2024 release with improved performance.
Anthropic's new Mythos-class flagship and the most capable coding model anyone can use — 80.3% SWE-Bench Pro, an 11-point jump over Opus 4.8. 1M context, 128K output, native parallel subagents. Released June 9, 2026.
Verdict
New global #1 — 80.3% SWE-Bench Pro, the most capable model generally available.
Quality score
98%
Pricing
$10.00/1M in
$50.00/1M out
Speed
Deliberate
2/5 speed
Context
1M tokens
Launched June 9, 2026 as the public, Mythos-class release. Available on the Claude API, Microsoft Foundry, and Google Vertex AI. Free for all users until June 22, 2026. Same underlying model as Claude Mythos 5, with safeguards that block specific high-risk cyber responses.
Coding leaderSWE-Bench Pro #1Mythos-classParallel subagentsAgenticLong contextPremiumNew
Best for
The hardest coding tasks, autonomous multi-step agents, and frontier-grade reasoning
Anthropic's September 1, 2026 frontier release and the new capability ceiling for coding, agents, and scientific work. Base pricing is unchanged at $10/$50, but cache reads dropped 75% to $0.25/1M — roughly 25% cheaper on typical workloads and up to 45% cheaper on agentic ones. 1M context, 128K output, adaptive thinking always on.
Verdict
New frontier leader — better than Fable 5 on every published benchmark, and cheaper to run.
Quality score
98%
Pricing
$10.00/1M in
$50.00/1M out
Speed
Deliberate
2/5 speed
Context
1M tokens
Released September 1, 2026 alongside Claude Mythos 5.1, the first update to the Mythos-class line since Fable 5 on June 9. API ID claude-fable-5-1; generally available on the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry. Published launch numbers (Fable 5.1 / Fable 5 / Opus 5 / GPT-5.6 Sol): Terminal-Bench-Science 0.1 52.6 / 24.7 / 29.0 / 22.4; Terminal-Bench 4.0 55.8 / 42.0 / 52.3 / 37.3; CursorBench 3.2.0 73.4 / 70.5 / 70.0 / 67.2; AutomationBench 31.4 / 17.1 / 26.9 / 19.6; OSWorld 2.0 strict 41.7 / 36.1 / 39.6; Humanity's Last Exam (no tools) 60.9 / 57.8 / 56.6; GDPval-AA v2 1853 / 1723 / 1824 / 1711. GDPval-AA v2 is rescaled from the v1 numbers quoted on the Fable 5 page and is not directly comparable to them.
Pricing moves, ranking shifts, and capability updates.
New ModelMar 27, 2026
Anthropic: Claude 3.7 Sonnet (thinking) — added to UseRightAI
Anthropic: Claude 3.7 Sonnet (thinking) (Anthropic) is now indexed. The most transparent reasoning model on the market — ideal when you need to see and trust the thought process, not just the answer.
Claude 3.7 Sonnet (thinking) costs $3 per million input tokens and $15 per million output tokens on the API. A month of 10M input and 2M output tokens runs about $60.00 at list price, before any batch or caching discounts.
What is Claude 3.7 Sonnet (thinking) best for?
Claude 3.7 Sonnet (thinking) is best for tackling complex coding challenges, mathematical proofs, and multi-step logical problems where visible reasoning and higher accuracy matter more than speed.. It is a strong fit when that workflow matters more than the tradeoffs around balanced pricing and deliberate speed.
When should I avoid Claude 3.7 Sonnet (thinking)?
You need fast, high-volume responses or are doing straightforward writing, summarization, or conversational tasks where extended thinking adds cost and latency without meaningful quality gains.
What is a cheaper alternative to Claude 3.7 Sonnet (thinking)?
GPT-5.1-Codex-Max (OpenAI) at $1.25/1M/1M input against Claude 3.7 Sonnet (thinking)'s $3.00/1M/1M — roughly 38% less per token all in. The strongest choice for serious software engineering work, provided you can absorb the output-side pricing. Compare it first if Claude 3.7 Sonnet (thinking)'s pricing is the thing stopping you.
What is a faster alternative to Claude 3.7 Sonnet (thinking)?
Claude 3.5 Sonnet — balanced against Claude 3.7 Sonnet (thinking)'s deliberate, with 200k tokens of context. Worth the swap when response time is what your users notice rather than the last few points of reasoning depth.
Newsletter
Get notified when Claude 3.7 Sonnet (thinking) pricing changes
We track pricing daily. When this model drops or spikes, you'll know first.
No spam. Useful updates only. Affiliate disclosures always clearly labeled.