https://sigmabench.com/ 2026-03-06 weekly 1.0 https://sigmabench.com/methodology/ 2026-01-21 monthly 0.4 https://sigmabench.com/company/ 2026-01-27 monthly 0.4 https://sigmabench.com/latest/ 2026-01-30 weekly 1.0 https://sigmabench.com/sample-report/ 2026-01-21 weekly 1.0 https://sigmabench.com/book-benchmark-call/ 2026-01-21 monthly 0.4 https://sigmabench.com/share/leaderboard/ 2026-01-30 weekly 1.0 https://sigmabench.com/blog/opencode-kimi-k-2-5-open-source-open-weight-agentic-coding-is-here/ 2026-01-30 monthly 0.4 https://sigmabench.com/assets/images/blog/2026/opencode-kimi-k-2-5-open-source-open-weight-agentic-coding-is-here.png OpenCode + Kimi K2.5 is fast, cheap, and surprisingly competitive. Our benchmarks show how close open-weight agentic coding is to the best proprietary models and where it already wins. https://sigmabench.com/blog/opencode-vs-codex-cli-on-gpt-5-1-codex-mini-and-5-2-codex/ 2026-01-29 monthly 0.4 https://sigmabench.com/assets/images/blog/2026/opencode-vs-codex-cli-on-gpt-5-1-codex-mini-and-5-2-codex.png OpenCode vs Codex CLI on Sigmabench: same Codex models, similar output quality, but Codex CLI's faster runtimes give it a consistent edge on the Sigmabench leaderboard. https://sigmabench.com/blog/opencode-vs-gemini-cli-on-gemini-3-flash-and-pro/ 2026-01-27 monthly 0.4 https://sigmabench.com/assets/images/blog/2026/opencode-vs-gemini-cli-on-gemini-3-flash-and-pro.png OpenCode outperforms Gemini CLI while using the same Gemini 3 models. We benchmark them head-to-head on Gemini 3 Flash Preview and Gemini 3 Pro Preview across Sigmascore, Accuracy, Consistency, and Speed.. https://sigmabench.com/blog/codex-cli-gpt-5-2-codex-most-accurate-ranks-3-sigmabench/ 2026-01-21 monthly 0.4 https://sigmabench.com/assets/images/blog/2026/codex-cli-gpt-5-2-codex-most-accurate-ranks-3-sigmabench.png GPT-5.2 Codex is the new flagship agentic model for Codex CLI. We benchmark it against the former flagship, GPT-5.1 Codex Max, and the legacy GPT-5.1 Codex. https://sigmabench.com/blog/the-origin-story-of-sigmabench/ 2026-01-19 monthly 0.4 https://sigmabench.com/assets/images/blog/2026/the-origin-story-of-sigmabench.png 2025 was the year autocomplete died. Agents took over. Inside complex enterprise codebases, the story was different. The code pushed back and often won. https://sigmabench.com/blog/gemini-3-flash-preview-beats-pro-on-sigmabench/ 2025-12-19 monthly 0.4 https://sigmabench.com/assets/images/blog/2025/gemini-3-flash-preview-beats-pro-on-sigmabench.png Gemini 3 Flash Preview beats Gemini 2.5 Flash and Gemini 3 Pro Preview across every Sigmabench metric for Gemini CLI users — ranking #2 overall. https://sigmabench.com/blog/there-is-no-best-agent-only-the-best-agent-for-your-codebase/ 2025-12-17 monthly 0.4 https://sigmabench.com/assets/images/blog/2025/there-is-no-best-agent-only-the-best-agent-for-your-codebase.png Claude, Cursor, Gemini, and Codex benchmarked on individual codebases. The insight: there is no best agent, only the best for your codebase. https://sigmabench.com/blog/cursor-composer-1-is-4x-faster-and-remains-number-1-overall/ 2025-12-17 monthly 0.4 https://sigmabench.com/assets/images/blog/2025/cursor-composer-1-is-4x-faster-and-remains-number-1-overall.png Cursor Composer-1 is 4x faster than the top Claude, Gemini & Codex models — ranking #1 on Sigmabench. https://sigmabench.com/blog/sigmabench-gains-soc2-type-1-and-2-accreditation/ 2025-09-27 monthly 0.4 https://sigmabench.com/assets/images/blog/2025/sigmabench-gains-soc2-type-1-and-2-accreditation.png Sigmabench has achieved SOC2 Type 2 attestation. Security guides everything we build — this audit makes that commitment visible.