DeepSWE

All models run on mini-swe-agent harness. Refer to DeepSWE for more details.

1. 1GPT-5.5[xhigh]

   70%±3%
   x
2. 2Claude Opus 4.8[max]

   58%±2%
   x
3. 3GPT-5.4[xhigh]

   56%±2%
   x
4. 4Claude Opus 4.7[max]

   54%±5%
   x

Pricing

Per-million-token pricing, prompt-cache rates, and long-context tiers for every model behind the Revolt endpoint.

OpenAIClaude

Model	Context	Input	Output	Cached
GPT-5.5	≤ 272K	$5	$30	$0.50
	> 272K	$10	$45	—
GPT-5.4	≤ 272K	$2.50	$15	$0.25
	> 272K	$5	$22.50	—