{"source":"BenchGecko","url":"https://benchgecko.ai/model/deepseek-v3-2-speciale","as_of":"2026-05-03","license":"BenchGecko collected data (prices, provider offers, Gecko Tests) is CC BY 4.0 · https://creativecommons.org/licenses/by/4.0/. Benchmark scores belong to their original publishers (see sources) and are aggregated with attribution. Attribution required: \"Source: BenchGecko\" with a link.","attribution":"Source: BenchGecko · https://benchgecko.ai/model/deepseek-v3-2-speciale","cite":"DeepSeek V3.2 Speciale · benchmarks, pricing and providers. BenchGecko, data as of 2026-05-03. https://benchgecko.ai/model/deepseek-v3-2-speciale","slug":"deepseek-v3-2-speciale","name":"DeepSeek V3.2 Speciale","provider":{"name":"DeepSeek","slug":"deepseek"},"model_type":"text","release_date":"2025-12-01","is_open_source":true,"status":"active","context_window":163840,"description":"DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance. It builds on DeepSeek Sparse Attention (DSA) for efficient long-context processing, then scales post-training reinforcement learning...","benchgecko_score":{"value":95.2,"rank":5,"of":312,"method":"Normalized average of public benchmark scores","method_url":"https://benchgecko.ai/methodology"},"avg_score":95.2,"list_price":{"input_usd_per_m":0.4,"output_usd_per_m":1.2,"source":"OpenRouter models API (checked daily)","as_of":"2026-10-05"},"pricing":{"input":0.4,"output":1.2},"providers":[],"cheapest_provider":null,"providers_source":null,"scores":[{"benchmark":"OpenCompass — AIME2025","benchmark_slug":"oc-aime2025","score":96,"unit":"%","category":"math","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-aime2025"},{"benchmark":"OpenCompass — IFEval","benchmark_slug":"oc-ifeval","score":91.7,"unit":"%","category":"language","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-ifeval"},{"benchmark":"OpenCompass — GPQA-Diamond","benchmark_slug":"oc-gpqa-diamond","score":86.7,"unit":"%","category":"knowledge","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-gpqa-diamond"},{"benchmark":"OpenCompass — MMLU-Pro","benchmark_slug":"oc-mmlu-pro","score":85.5,"unit":"%","category":"knowledge","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-mmlu-pro"},{"benchmark":"OpenCompass — LiveCodeBenchV6","benchmark_slug":"oc-livecodebenchv6","score":80.9,"unit":"%","category":"coding","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-livecodebenchv6"},{"benchmark":"Artificial Analysis — Coding Index","benchmark_slug":"aa-coding-index","score":37.89,"unit":"index","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-coding-index"},{"benchmark":"Artificial Analysis — Quality Index","benchmark_slug":"aa-quality-index","score":29.43,"unit":"index","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-quality-index"},{"benchmark":"OpenCompass — HLE","benchmark_slug":"oc-hle","score":28.6,"unit":"%","category":"knowledge","source":"OpenCompass","source_url":"https://rank.opencompass.org.cn","benchmark_url":"https://benchgecko.ai/benchmark/oc-hle"},{"benchmark":"Artificial Analysis — Agentic Index","benchmark_slug":"aa-agentic-index","score":0,"unit":"index","category":"speed","source":"Artificial Analysis","source_url":"https://artificialanalysis.ai","benchmark_url":"https://benchgecko.ai/benchmark/aa-agentic-index"}],"gecko_tests":null,"links":{"page":"https://benchgecko.ai/model/deepseek-v3-2-speciale","json":"https://benchgecko.ai/api/v1/models/deepseek-v3-2-speciale","price_history":"https://benchgecko.ai/api/v1/price-history/deepseek-v3-2-speciale","provider_prices":null}}