Loading model details, price history, and benchmarks…
Loading model details, price history, and benchmarks…
azure
Claude Sonnet 5 is designed for fast coding, tool use, and multi-step agent workflows. It supports adaptive thinking, text and image input, and a million-token context window.
From provider documentation. Reviewed September 18, 2026.
Compare exact models and providers before you fill up.
azure
Loading estimate…
OpenAI
Loading estimate…
Estimates use tracked list rates. They are not actual charges or a billing commitment. Taxes, tool fees, negotiated rates and provider-specific limits may differ. Provider sign-in is required to buy credits.
Anthropic made the introductory $2 input / $10 output rates permanent. Older benchmark charts may still calculate cost using $3 / $15.
Hover, tap, or use the day slider to inspect prices.
23 of 23 days have observed prices. Gaps are not assumed unchanged. Successful unchanged-content checks count only when an earlier matching source snapshot proves the model was present. All dates are UTC.
| Date (UTC) | Input | Output | Evidence |
|---|---|---|---|
| 2026-09-08 | $2 | $10 | unverified live observation |
| 2026-09-09 | $2 | $10 | unverified live observation |
| 2026-09-10 | $2 | $10 | unverified live observation |
| 2026-09-11 | $2 | $10 | unverified live observation |
| 2026-09-12 | $2 | $10 | unverified live observation |
| 2026-09-13 | $2 | $10 | unverified live observation |
| 2026-09-14 | $2 | $10 | unverified live observation |
| 2026-09-15 | $2 | $10 | unverified live observation |
| 2026-09-16 | $2 | $10 | unverified live observation |
| 2026-09-17 | $2 | $10 | unverified live observation |
| 2026-09-18 | $2 | $10 | unverified live observation |
| 2026-09-19 | $2 | $10 | unverified live observation |
| 2026-09-20 | $2 | $10 | unverified live observation |
| 2026-09-21 | $2 | $10 | unverified live observation |
| 2026-09-22 | $2 | $10 | unverified live observation |
| 2026-09-23 | $2 | $10 | unverified live observation |
| 2026-09-24 | $2 | $10 | unverified live observation |
| 2026-09-25 | $2 | $10 | unverified live observation |
| 2026-09-26 | $2 | $10 | unverified live observation |
| 2026-09-27 | $2 | $10 | unverified live observation |
| 2026-09-28 | $2 | $10 | unverified live observation |
| 2026-09-29 | $2 | $10 | unverified live observation |
| 2026-09-30 | $2 | $10 | unverified live observation |
Each source has its own history. We don’t join conflicting sources into a single price trend. Standard input prices exclude cache discounts.
These are the earliest and latest successful checks we can prove from our retained polling records. Reconstructing an older price never moves these timestamps backward.
A baseline of these monitoring records was preserved on . The current checks below may be newer.
| Source | First proven live check (UTC) | Latest proven live check (UTC) |
|---|---|---|
| models-dev |
A reconstruction from dated sources. The effective date, publication date, and date we added the evidence are recorded separately. Coverage is incomplete; an absent event does not prove nothing changed.
Earlier launches and price changes have not been verified for this listing. The price chart above uses our recorded observations independently; daily collection does not wait for historical research.
Results depend on reasoning effort, tools, and evaluation setup. Check the original report before comparing scores across models.
| Benchmark | Score | Evaluation setup | Source |
|---|---|---|---|
| SWE-bench Pro | 63.2% | Launch benchmark table; see system card for harness and effort | Provider reportPublished 2026-06-30 |
| Terminal-Bench 2.1 | 80.4% | Launch benchmark table; see system card for harness and effort | Provider reportPublished 2026-06-30 |
| Humanity’s Last Exam | 43.2% | No tools; launch benchmark table | Provider reportPublished 2026-06-30 |
| Humanity’s Last Exam | 57.4% | With tools; launch benchmark table | Provider reportPublished 2026-06-30 |
| OSWorld-Verified | 81.2% | Launch benchmark table; different evaluation from OSWorld 2.0 | Provider reportPublished 2026-06-30 |
Provider-reported, not independently verified by TokenPylon. Scores and settings are quoted from the linked tables; these are not a live leaderboard.
| Benchmark | Recorded score | Source | Date |
|---|---|---|---|
| lmarena-text-elohigh; leaderboard=text_style_control; category=overall; published=2026-09-25claude-sonnet-5-high [leaderboard=text_style_control; category=overall; published=2026-09-25] |
| 1,461.526Elo |
| lmarena |
| 2026-09-25 |
| simple-benchscore=AVG@5Claude Sonnet 5 [score=AVG@5] | 0.606fraction | simple-bench | 2026-09-23 |
| lmarena-text-elohigh; leaderboard=text_style_control; category=overall; published=2026-09-13claude-sonnet-5-high [leaderboard=text_style_control; category=overall; published=2026-09-13] | 1,461.207Elo | lmarena | 2026-09-13 |
| livebench-averagexhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.7604fraction | livebench | 2026-06-25 |
| livebench-reasoningxhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.8869fraction | livebench | 2026-06-25 |
| livebench-ifxhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.6386fraction | livebench | 2026-06-25 |
| livebench-agentic-codingxhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.5939fraction | livebench | 2026-06-25 |
| livebench-mathematicsxhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.9294fraction | livebench | 2026-06-25 |
| livebench-languagexhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.7497fraction | livebench | 2026-06-25 |
| livebench-codingxhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.8068fraction | livebench | 2026-06-25 |
| livebench-data-analysisxhigh-effort; release=2026-06-25; score=mean of task means per categoryclaude-sonnet-5-xhigh-effort [release=2026-06-25; score=mean of task means per category] | 0.7174fraction | livebench | 2026-06-25 |
Recorded scores use the source’s stored scale; percentage accuracies may be fractions from 0 to 1. Evaluation settings are not consistently available in the tracked feed. Provider-self-reported results have not been independently verified by TokenPylon.
Published by the model’s provider; not independent verification by TokenPylon.