Benchmark score change log
A published score on Newly Stack can only change together with an entry in this log that says why. The database refuses any score change without one. Corrections after a dispute appear here too.
| Date | Model | Suite | Score | Reason |
|---|---|---|---|---|
| October 25, 2026 | Command A+ | ns-id-v1 | first published: 87.83 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Mistral Small 4 | ns-id-v1 | first published: 84.50 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Mistral Medium 3.5 | ns-id-v1 | first published: 89.60 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Qwen3.8 Max (0902) | ns-id-v1 | first published: 97.55 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Kimi K3.1 Pro | ns-id-v1 | first published: 99.26 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | DeepSeek V4.1 Flash | ns-id-v1 | first published: 97.13 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | DeepSeek V4 Pro | ns-id-v1 | first published: 96.80 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Grok 4.7 | ns-id-v1 | first published: 97.98 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Perplexity Sonar | ns-id-v1 | first published: 97.11 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Llama 4 Maverick | ns-id-v1 | first published: 88.15 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Claude Haiku 5.5 | ns-id-v1 | first published: 93.19 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Claude Sonnet 5.5 | ns-id-v1 | first published: 97.64 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Claude Opus 5.5 | ns-id-v1 | first published: 98.97 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Gemini 3.5 Flash-Lite | ns-id-v1 | first published: 90.69 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Gemini 3.8 Flash | ns-id-v1 | first published: 93.00 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | Gemini 4 Argon | ns-id-v1 | first published: 94.58 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | GPT-6 Luna | ns-id-v1 | first published: 97.40 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | GPT-6.1 Sol | ns-id-v1 | first published: 98.46 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | GPT-6 Astra | ns-id-v1 | first published: 97.99 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
| October 25, 2026 | ChatGPT (chat-latest) | ns-id-v1 | first published: 96.40 | First publication of NS-ID v1 (run of 25 October 2026, model-graded), reviewed before launch. |
Scoring corrections before publication
Before the nsid-v1-2026-10 results went public we corrected the automatic scorer and recomputed every automatic score from the stored raw outputs, without new API calls. The task scores that changed are listed below. Judge scores were not touched.
| Model | Task | Before | After |
|---|---|---|---|
| kimi-k3-1-pro | ins-08 | 50 | 100 |
| claude-opus-5-5 | ins-08 | 50 | 100 |
| grok-4-7 | ins-08 | 50 | 100 |
| gpt-6-1-sol | kno-08 | 0 | 100 |
| gpt-6-astra | kno-08 | 0 | 100 |
| deepseek-flash | ins-08 | 50 | 100 |
| perplexity-sonar | kno-08 | 0 | 100 |
| claude-sonnet-5-5 | ins-08 | 50 | 100 |
| claude-sonnet-5-5 | kno-03 | 0 | 100 |
| gpt-6-luna | ins-08 | 50 | 100 |
| gpt-6-luna | kno-08 | 0 | 100 |