Leaderboard

SWE-Atlas Refactoring

11 models scored · metric: score. Prices are the cheapest listed across all serving providers. Cost/run and Pts/$ are priced for the long context token profile — a whole codebase or filing per request — the case tiered pricing exists for.

A benchmark is priced for the workload it exercises, not a fixed input:output blend: an agent loop and a single hard question bill very differently on the same rate card. Profile: 250k in · 2k out · 30% cache hit · 256k context.

Scores are facts as publicly reported by labs (see report links below and on each model page) and aggregated via models.dev. SWE-Atlas Refactoring is maintained by its own project — Model Pulse is not affiliated with or endorsed by it.

Best value · top 5 by pts per dollar · long context profile

#
🥇	Claude Opus 4.7anthropic	48.57	$4	$20	$0.895	54.3	1.05M
🥈	GPT-5.5openai	44.79	$0.188	$1.13	$0.054	832	1.05M
🥉	GPT-5.4openai	44.29	$0.75	$6	$0.168	264	1.05M
4	GPT-5.3 Codexopenai	42.38	$1.60	$13	$0.358	118	400K
5	Claude Opus 4.6anthropic	35.58	$4	$20	$0.895	39.8	1M
6	Gemini 3.1 Pro Previewgoogle	33.81	$1	$6	$0.199	170	1.05M
7	Claude Sonnet 4.6anthropic	32.21	$0.9	$5.55	$0.198	163	1M
8	GLM-5zhipuai	24.24	$0.35	$1.40	$0.086	282	205K
9	Kimi K2.5moonshotai	20.95	$0.3	$1.90	$0.075	279	262K
10	MiniMax-M2.5minimax	19.52	$0.22	$0.88	$0.062	314	229K
11	Gemini 3 Flash Previewgoogle	10	$0.07	$0.43	$0.020	497	1.05M

Score reports: labs.scale.com — full source URLs are linked on each model page.

Cheapest scorer: Gemini 3 Flash Preview at $0.07 input /M.