UnifyBench

RankingFrontierSWE v2

FrontierSWE v2

Data updated 13 Oct 2026

Bucket
Supporting evidence
Unit
percent
Direction
Higher is better
Version
2
Display harness
Proximal FrontierSWE v2 · Proximus · unspecified
Board
https://www.frontierswe.com/

Proximal FrontierSWE v2 uses mean@5 across 34 long-horizon coding tasks, nominal five trials per task and a 20-hour task budget, with Proximus and independent verifiers. Actual task trial counts can vary. Published headline scores are retained when task-summary means differ. Named per-model effort settings are unspecified. Historical provider claims remain secondary source evidence. These rows stay out of weighted Overall.

Compare published benchmark results with category weights →

Models

Results are ordered by score in the display harness, followed by models with no result. Other harnesses are listed separately below. Missing results remain unknown.

226–250 of 469 entries

ModelScoreEvidenceSource-recorded date
Hunyuan-1.8B-InstructTencent———
Hunyuan-4B-InstructTencent———
Hunyuan-7B-InstructTencent———
Hunyuan-7B-Instruct-0124Tencent———
Hunyuan-A13B-InstructTencent———
Hy3Tencent———
Hy3-previewTencent———
Hy4-previewTencent———
K EXAONE 2.0 750B A37BLG AI Research———
K EXAONE 236B A23BLG AI Research———
Kimi-Dev-72BMoonshot———
Kimi-K2.6Moonshot———
Kimi-K2.7-CodeMoonshot———
Kimi-Linear-48B-A3B-InstructMoonshot———
Kimi-VL-A3B-InstructMoonshot———
Kimi-VL-A3B-Thinking-2506Moonshot———
Leanstral 1.5Mistral———
Llama 4 MaverickMeta———
Llama-3.1-405B-InstructMeta———
Llama-3.1-70B-InstructMeta———
Llama-3.1-8B-InstructMeta———
Llama-3.2-11B-Vision-InstructMeta———
Llama-3.2-1B-InstructMeta———
Llama-3.2-3B-InstructMeta———
Llama-3.2-90B-Vision-InstructMeta———