Claude Sonnet 4.5 Thinking
| Released | 2025-09-29 |
|---|---|
| Price per 1M | $3.00 in · $15.00 out |
| Overall score | 51.4 · rank 249 |
| Context window | 200,000 tokens |
| Provider | Anthropic |
| Weights | Proprietary |
| Configuration | Thinking |
Scores
| Dimension | Score | Rank | Benchmarks |
|---|---|---|---|
| Overall: Balanced | 51.4 | 249 | 20 |
| Overall: Coding | 50.9 | 248 | 12 |
| Overall: Reasoning | 51.8 | 238 | 13 |
| Overall: Agentic work | 52.2 | 245 | 11 |
| Overall: Frontend | 48.2 | 245 | 5 |
| Coding | 51.2 | 236 | 2 |
| Agents | 52.5 | 180 | 2 |
| Reasoning | 52.4 | 229 | 7 |
| Math | 51.3 | 166 | 4 |
| Knowledge | 51.4 | 114 | 2 |
| Frontend | 46.3 | 141 | 1 |
| Multimodal | 50.7 | 110 | 2 |
| Value vs price | −3.9 | 299 |
Evidence
Coding 51.2 ±3.3
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SWE-bench Verified (Vals run) | 70.0% | Vals AI | 2026-09-01 |
| LiveCodeBench (Vals run) | 73.0% | Vals AI | 2026-09-01 |
Agents 52.5 ±2.9
| Benchmark | Result | Source | Date |
|---|---|---|---|
| METR task suite | 67.4% | Epoch AI | 2026-11-11 |
| DeepResearch Bench | 52.6% | Epoch AI | 2026-11-11 |
Reasoning 52.4 ±1.7
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Extended NYT Connections | 37.3% | Lech Mazur | 2026-11-10 |
| ARC-AGI-2 | 13.6% | Epoch AI | 2026-11-11 |
| ARC-AGI-2 (ARC Prize verified) | 13.6% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 (ARC Prize verified) | 63.7% | ARC Prize | 2026-11-11 |
| ARC-AGI-1 | 63.7% | Epoch AI | 2026-11-11 |
| SimpleBench | 54.3% | Epoch AI | 2026-11-11 |
| GPQA Diamond | 82.3% | Epoch AI | 2025-10-28 |
| GPQA Diamond (Vals run) | 81.6% | Vals AI | 2026-09-01 |
| Chess Puzzles | 10.0% | Epoch AI | 2026-10-13 |
| Mystery Game Puzzles | 17.0% | Epoch AI | 2026-07-25 |
Math 51.3 ±1.2
| Benchmark | Result | Source | Date |
|---|---|---|---|
| FrontierMath Tiers 1-3 | 23.9% | Epoch AI | 2026-06-11 |
| MATH Level 5 | 97.7% | Epoch AI | 2025-10-21 |
| FrontierMath Tier 4 | 2.4% | Epoch AI | 2026-06-11 |
| OTIS Mock AIME 2024-2025 | 77.8% | Epoch AI | 2025-10-28 |
| FrontierMath Tiers 1-3 (2025 set) | 15.2% | Epoch AI | 2025-11-16 |
| FrontierMath Tier 4 (2025 set) | 4.2% | Epoch AI | 2025-10-22 |
Knowledge 51.4 ±2.5
| Benchmark | Result | Source | Date |
|---|---|---|---|
| SimpleQA Verified | 30.7% | Epoch AI | 2026-08-27 |
| MMLU-Pro (Vals run) | 87.4% | Vals AI | 2026-09-01 |
Frontend 46.3 ±0.8
| Benchmark | Result | Source | Date |
|---|---|---|---|
| Vibe Code Bench v1.1 | 22.6% | Vals AI | 2026-11-09 |
Multimodal 50.7 ±2.4
| Benchmark | Result | Source | Date |
|---|---|---|---|
| MMMU-Pro (Vals run) | 79.3% | Vals AI | 2026-09-01 |
| VPCT | 39.8% | Epoch AI | 2026-11-11 |
Names matched to this model
claude-sonnet-4-5-20250929-thinking