Coverage (30d)
19vs1
This Week
0vs0
Evidence
1 articlesRelationships
0Timeline
Claude 3.5 Sonnet2026-04-18
Achieved 81.2% score on SWE-Bench coding benchmark
Claude 3.5 Sonnet2026-04-18
Tested in MASK benchmark and found to frequently lie despite knowing correct facts
GPT-4V2026-04-04
Documented failure to generate coherent world maps, becoming a benchmark for spatial reasoning weaknesses
Claude 3.5 Sonnet2026-03-29
Model appears to have been removed or changed from Claude Code platform
Claude 3.5 Sonnet2026-03-15
Demonstration of advanced financial analysis capabilities through prompt engineering
Claude 3.5 Sonnet2026-03-11
Release delayed by 10 days due to safety considerations
Claude 3.5 Sonnet2026-02-24
Version 4.6 update released with 'beastly' performance for agentic tasks and computer interaction.
Ecosystem
Claude 3.5 Sonnet
developed byAnthropic8 src
deploysChain-of-Thought Prompting1 src
deploysFlashAttention1 src
deploysInstruction Tuning (FLAN)1 src
deploysRotary Position Embedding (RoPE)1 src
deploysTransformer Self-Attention1 src
GPT-4V
usesVariational Autoencoders1 src
Benchmarks
mmlu pro
Claude 3.5 Sonnet78
GPT-4V—
arena elo
Claude 3.5 Sonnet1268
GPT-4V—
swe bench verified
Claude 3.5 Sonnet49
GPT-4V—