Model card
Claude Sonnet 4.
AnthropicapiMultimodal LLMProprietary1 current SOTA
#1 on ThaiOCRBench
§ 01 · Benchmarks
Every benchmark Claude Sonnet 4 has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | CrossCodeEval | Computer Code · Code Completion | exact-match | 44.5% | #1 | — | source ↗ |
| 02 | ThaiOCRBench | Computer Vision · Optical Character Recognition | ted-score | 0.8% | #1 | — | source ↗ |
| 03 | TransCoder (GeeksForGeeks) | Computer Code · Code Translation | computational-accuracy | 89.4% | #1 | — | source ↗ |
| 04 | MBPP | Computer Code · Code Generation | pass@1 | 89.6% | #6 | — | source ↗ |
| 05 | GSM8K | Reasoning · Mathematical Reasoning | accuracy | 97.8% | #12 | — | source ↗ |
| 06 | MMLU | Reasoning · Commonsense Reasoning | accuracy | 90.1% | #14 | — | source ↗ |
| 07 | HumanEval | Computer Code · Code Generation | pass@1 | 90.6% | #16 | — | source ↗ |
| 08 | SWE-Bench Verified | Computer Code · Code Generation | resolve-rate | 72.7% | #16 | — | source ↗ |
| 09 | OCRBench v2 | Computer Vision · General OCR Capabilities | overall-en-private | 42.4% | #17 | 2025-05-22 | source ↗ |
| 10 | LiveCodeBench | Computer Code · Code Generation | pass@1 | 52.8% | #18 | 2024-03-12 | source ↗ |
| 11 | GPQA | Reasoning · Multi-step Reasoning | accuracy | 70.0% | #18 | — | source ↗ |
| 12 | MATH | Reasoning · Mathematical Reasoning | accuracy | 88.9% | #21 | — | source ↗ |
| 13 | SWE-bench Verified | Agentic AI · SWE-bench | resolve-rate | 72.7% | #32 | — | source ↗ |
| 14 | PLCC | Natural Language Processing · Polish Cultural Competency | history | 81.0% | #60 | — | source ↗ |
| 15 | PLCC | Natural Language Processing · Polish Cultural Competency | culture-and-tradition | 72.0% | #64 | — | source ↗ |
| 16 | PLCC | Natural Language Processing · Polish Cultural Competency | vocabulary | 61.0% | #66 | — | source ↗ |
| 17 | PLCC | Natural Language Processing · Polish Cultural Competency | average | 68.2% | #69 | — | source ↗ |
| 18 | PLCC | Natural Language Processing · Polish Cultural Competency | geography | 77.0% | #71 | — | source ↗ |
| 19 | PLCC | Natural Language Processing · Polish Cultural Competency | grammar | 63.0% | #74 | — | source ↗ |
| 20 | PLCC | Natural Language Processing · Polish Cultural Competency | art-and-entertainment | 55.0% | #77 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area
Where Claude Sonnet 4 actually performs.
§ 03 · Papers
1 paper with results for Claude Sonnet 4.
- 2024-03-12· Computer Code· 1 result
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code
§ 04 · Related models
Other Anthropic models scored on Codesota.
Claude Opus 4
Undisclosed params · 13 results · 2 SOTA
Claude Opus 4.5
3 results · 2 SOTA
Claude Sonnet 5
Undisclosed params · 2 results · 2 SOTA
Claude Mythos Preview
1 result · 1 SOTA
Claude 3.5 Sonnet
Undisclosed params · 27 results
Claude Opus 4.5
Undisclosed params · 13 results
Claude 3.7 Sonnet
10 results
Claude 3 Opus
5 results
§ 05 · Sources & freshness
Where these numbers come from.
sdadas/PLCC
7
results
official-model-card
4
results
anthropic-model-card
4
results
alphaxiv-leaderboard
1
result
anthropic-blog
1
result
ocrbench-v2-leaderboard
1
result
official-leaderboard
1
result
editorial
1
result
16 of 20 rows marked verified. · first result 2024-03-12, latest 2025-05-22.