western lab · United States
Anthropic
Safety-focused lab behind the Claude model family, known for constitutional AI, interpretability research and enterprise-heavy distribution. Backed by Amazon and Google.
- Founded
- 2021
- Models tracked
- 43
- Valuation
- $183B
- Country
- United States
est. · as of Aug 2026
Founders
Dario Amodei · Daniela Amodei · Tom Brown · Jared Kaplan · Chris Olah
Model Benchmarks
updated just nowartificial analysis ↗| Model | Released | Intelligence | Coding | GPQA | HLE | SciCode | Terminal-Bench Hard | Tau2 | LCR | IFBench | Speed | TTFT | $/1M in | $/1M out | $/1M blended |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Claude Opus 5 (Adaptive Reasoning, Max Effort) | 2026-07-24 | 60.7 | 78 | 93.2% | 52.6% | 55.7% | — | — | 70% | — | 61.1 t/s | 27.78s | $5 | $25 | $10 |
| Claude Opus 5 (Adaptive Reasoning, Xhigh Effort) | 2026-07-24 | 60.1 | 77 | 93.7% | 52.5% | 55% | — | — | 68.7% | — | 58.2 t/s | 21.88s | $5 | $25 | $10 |
| Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) | 2026-06-09 | 59.9 | 76.5 | 92.6% | 53.3% | 60.2% | 62.9% | 98.5% | 70% | 63.5% | 70.7 t/s | 86.76s | $10 | $50 | $20 |
| Claude Opus 5 (Adaptive Reasoning, High Effort) | 2026-07-24 | 58.9 | 76.5 | 93.7% | 50.7% | 54.3% | — | — | 67% | — | 58.6 t/s | 12.1s | $5 | $25 | $10 |
| Claude Opus 5 (Adaptive Reasoning, Medium Effort) | 2026-07-24 | 56.3 | 74.3 | 91.9% | 49.2% | 50.7% | — | — | 69% | — | 59 t/s | 9.22s | $5 | $25 | $10 |
| Claude Opus 4.8 (Adaptive Reasoning, Max Effort) | 2026-05-28 | 55.7 | 74.3 | 92% | 45.7% | 53.5% | 58.3% | 94.4% | 67.7% | 62.2% | 0 t/s | 0s | $5 | $25 | $10 |
| Claude Opus 4.7 (Adaptive Reasoning, Max Effort) | 2026-04-16 | 53.5 | 73.6 | 91.4% | 39.6% | 54.5% | 51.5% | 88.6% | 70.3% | 58.6% | 0 t/s | 0s | $5 | $25 | $10 |
| Claude Sonnet 5 (Adaptive Reasoning, Max Effort) | 2026-06-30 | 53.4 | 71.5 | 91.1% | 39.6% | 53.6% | — | — | 70.7% | — | 96.7 t/s | 138.52s | $2 | $10 | $4 |
| Claude Opus 5 (Adaptive Reasoning, Low Effort) | 2026-07-24 | 50.6 | 66.9 | 88.9% | 41.3% | 48% | — | — | 69.3% | — | 57.2 t/s | 3.16s | $5 | $25 | $10 |
| Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | 2026-02-17 | 47.2 | 63 | 87.5% | 30% | 46.8% | 53% | 75.7% | 70.7% | 56.6% | 0 t/s | 0s | $3 | $15 | $6 |
Indices on Artificial Analysis 0-100 scale · eval scores are raw accuracy · “—” means the metric is not published for that model
Newsroom
source ↗- 2026-07-24Introducing Claude Opus 5Model Launch
- 2026-07-09Inviting hard questionsRelease
- 2026-06-30Redeploying Fable 5Model Launch
- 2026-06-30Introducing Claude Sonnet 5Model Launch
- 2026-08-04Mariano-Florentino (Tino) Cuéllar to join Anthropic as Chief Global Affairs OfficerRelease
- 2026-07-30Investigating three real-world incidents in our cybersecurity evaluationsRelease
- 2026-07-27Our position on open-weights modelsRelease
- 2026-07-27Cognizant and Anthropic expand their partnership to bring Claude to enterprise clientsRelease
Partners & Investors
disclosedFunding rounds
- Series B$580M
FTX / Sam Bankman-Fried · Apr 2022
- Series C$450M
Spark Capital · May 2023
- Strategic$4B total
Amazon · 2023-2024
- Strategic$3B total
Google · 2023-2025
- Series E$13B
ICONIQ, Fidelity, Lightspeed · Sep 2025
Partners
- Amazon Web ServicesCloud
- Google Cloud (TPU)Compute
- AWS Trainium (Project Rainier)Chips
- Palantir + AWS (gov)Enterprise
- Accenture / DeloitteEnterprise
Publicly announced rounds and partnerships only · as of Aug 2026