DeepSWE Updates

https://rss.xlit.app/deepswe DeepSWE Updates https://deepswe.datacurve.ai/ Reporting: Updated DeepSeek v4 Pro and DeepSeek v4 Flash results to account for adjusted pricing after DeepSeek's Aug 16 price change, which raised both models' rates and introduced peak and off-peak pricing. Costs use the peak rates; off-peak is half. Thu, 20 Aug 2026 22:00:00 GMT 7E17930D 2026-08-20T22:00:00Z https://deepswe.datacurve.ai/ Reporting: Updated GPT-5.6 Sol results to account for adjusted pricing after OpenAI's Aug 20 price cut (20% on input, 33% on output; promotional through at least Nov 21, 2026). Thu, 20 Aug 2026 22:00:00 GMT 146CE403 2026-08-20T22:00:00Z https://deepswe.datacurve.ai/ Reporting: Updated Gemini 3.6 Flash results to account for adjusted pricing after Google's Aug 13 price cut, which moved it to the same 50% introductory rate as Gemini 3.7 Flash through Dec 31, 2026. Thu, 13 Aug 2026 22:00:00 GMT FB94DCEE 2026-08-13T22:00:00Z https://deepswe.datacurve.ai/ Reporting: Corrected DeepSeek v4 Pro costs. DeepSeek's May 22 75% price cut was being applied to results that were already billed at the reduced rates. Thu, 13 Aug 2026 22:00:00 GMT 9B1B2A42 2026-08-13T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Gemini 3.7 Flash results. Wed, 12 Aug 2026 22:00:00 GMT 647DD3CA 2026-08-12T22:00:00Z https://deepswe.datacurve.ai/ Reporting: Re-ran Gemini 3.1 Pro, Gemini 3.5 Flash, and Gemini 3.6 Flash. A LiteLLM issue had double-counted tokens and overestimated their costs. Wed, 12 Aug 2026 22:00:00 GMT 6A5F160D 2026-08-12T22:00:00Z https://deepswe.datacurve.ai/ Models: Added DeepSeek v4 Pro results. Tue, 11 Aug 2026 22:00:00 GMT 08CE70A6 2026-08-11T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Grok 4.6 results. Tue, 11 Aug 2026 22:00:00 GMT A9291022 2026-08-11T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Muse Spark 1.2 results. Thu, 06 Aug 2026 22:00:00 GMT E5A5EC8C 2026-08-06T22:00:00Z https://deepswe.datacurve.ai/ Models: Added DeepSeek v4 Flash results. Wed, 05 Aug 2026 22:00:00 GMT 4D95628F 2026-08-05T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Qwen 3.8 Max results. Mon, 03 Aug 2026 22:00:00 GMT 9504077A 2026-08-03T22:00:00Z https://deepswe.datacurve.ai/ Reporting: Updated GPT-5.6 Terra and GPT-5.6 Luna results to account for adjusted pricing after OpenAI's Jul 30 rate cuts. Wed, 29 Jul 2026 22:00:00 GMT 511D5752 2026-07-29T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Claude Opus 5 results. Fri, 24 Jul 2026 22:00:00 GMT 235E5551 2026-07-24T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Gemini 3.6 Flash results. Tue, 21 Jul 2026 22:00:00 GMT E2995E4E 2026-07-21T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Kimi K3 results. Fri, 17 Jul 2026 22:00:00 GMT 9189F525 2026-07-17T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Grok 4.5 results. Wed, 15 Jul 2026 22:00:00 GMT BC3FAB16 2026-07-15T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Muse Spark 1.1 results. Mon, 13 Jul 2026 22:00:00 GMT BA906D24 2026-07-13T22:00:00Z https://deepswe.datacurve.ai/ Models: Added GPT-5.6 Sol, GPT-5.6 Terra, and GPT-5.6 Luna results. Thu, 09 Jul 2026 22:00:00 GMT 7E602C20 2026-07-09T22:00:00Z https://deepswe.datacurve.ai/ Models: Added Claude Sonnet 5 results. Wed, 01 Jul 2026 22:00:00 GMT 66A6AA19 2026-07-01T22:00:00Z https://deepswe.datacurve.ai/ Models: Added GLM 5.2 results. Sat, 20 Jun 2026 22:00:00 GMT F3F0BE48 2026-06-20T22:00:00Z https://deepswe.datacurve.ai/ Benchmark: DeepSWE v1.1 released: 113 tasks with isolated verification and structured test reports. Sun, 14 Jun 2026 22:00:00 GMT CBA8A8F1 2026-06-14T22:00:00Z https://deepswe.datacurve.ai/ Benchmark: Initial leaderboard: Claude Fable 5, Claude Opus 4.8, Claude Sonnet 4.6, Gemini 3.1 Pro, Gemini 3.5 Flash, GPT-5.4, GPT-5.5, and Kimi K2.7 Code. Sun, 14 Jun 2026 22:00:00 GMT 6008A5B1 2026-06-14T22:00:00Z