List your company

Agentic Coding on Terminal Bench 2.0

59.1Pass@1Claude Sonnet 4.6
21.03630.91840.850.682Mar 21, 2026Apr 1, 2026Apr 12, 2026Apr 23, 2026May 4, 2026May 15, 2026May 26, 2026

Evaluation results

MethodLinks
Claude Sonnet 4.6# Total Params=-, # Active Params=-2026.0559.1
DeepSeek-V4 Flash# Total Params=284B, # Active Params=13B2026.0556.9
Gemini 3-pro2026.0354.2
Qwen3.5# Total Params=397B, # Active Params=17B2026.0552.5
Qwen3.6# Total Params=35B, # Active Params=3B2026.0551.5
GPT-5.4 Nano2026.0546.3
LAGUNA M.1# Total Params=225B, # Active Params=23B2026.0545.8
Seed1.82026.0345.2
Gemma 4# Total Params=31B, # Active Params=31B2026.0542.9
Claude Sonnet-4.52026.0342.8
GLM-4.7# Total Params=355B, # Active Params=32B2026.0541
Qwen3.5# Total Params=35B, # Active Params=3B2026.0540.5
LAGUNA XS.2# Total Params=33.4B, # Active Params=3B2026.0535.7
GPT-5 High2026.0335.2
Gemini 2.5-pro2026.0332.6
Devstral 2# Total Params=123B, # Active Params=123B2026.0532.6
Claude Haiku 4.52026.0529.8
Devstral Small 2# Total Params=24B, # Active Params=24B2026.0522.5