| model,gateway,served_backend,model_revision,quantization,task_type,context_format,output_contract,metric_axis,k_star,k_ladder,cases,repetitions,provider_pinned,source_title,source_location,source_url,benchmark_code,contributor,status,notes | |
| gpt-5.5-pro,OpenRouter,OpenAI,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,8,0|1|2|3|5|8|13|full,6,1,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot,Reasoning effort low; 320-token output cap. | |
| gpt-5.5-pro,OpenRouter,OpenAI,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,3,0|1|2|3|5|8|13|full,6,1,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot,Reasoning effort low; 320-token output cap. | |
| gpt-5.5-pro,OpenRouter,OpenAI,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,1,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot,Reasoning effort low; 320-token output cap. | |
| gpt-5.5-pro,OpenRouter,OpenAI,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,1,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot,Reasoning effort low; 320-token output cap. | |
| gpt-5.5-pro,OpenRouter,OpenAI,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,13,0|1|2|3|5|8|13|full,6,1,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot,Reasoning effort low; 320-token output cap. | |
| claude-opus-4.8,OpenRouter,Anthropic,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,13,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| claude-opus-4.8,OpenRouter,Anthropic,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| claude-opus-4.8,OpenRouter,Anthropic,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| claude-opus-4.8,OpenRouter,Anthropic,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| claude-opus-4.8,OpenRouter,Anthropic,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,13,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| gpt-4o,OpenRouter,OpenAI,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| gpt-4o,OpenRouter,OpenAI,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,13,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| gpt-4o,OpenRouter,OpenAI,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| gpt-4o,OpenRouter,OpenAI,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| gpt-4o,OpenRouter,OpenAI,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.1-8b,OpenRouter,WandB,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.1-8b,OpenRouter,WandB,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.1-8b,OpenRouter,WandB,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.1-8b,OpenRouter,WandB,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,13,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.1-8b,OpenRouter,WandB,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,1,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.0-h-micro,OpenRouter,Cloudflare,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.0-h-micro,OpenRouter,Cloudflare,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.0-h-micro,OpenRouter,Cloudflare,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.0-h-micro,OpenRouter,Cloudflare,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| granite-4.0-h-micro,OpenRouter,Cloudflare,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| qwen-2.5-7b,OpenRouter,Together,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| qwen-2.5-7b,OpenRouter,Together,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,8,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| qwen-2.5-7b,OpenRouter,Together,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| qwen-2.5-7b,OpenRouter,Together,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,2,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| qwen-2.5-7b,OpenRouter,Together,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,3,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| llama-3.2-3b,OpenRouter,Cloudflare,,,factual,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,2,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| llama-3.2-3b,OpenRouter,Cloudflare,,,multi-hop,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,2,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| llama-3.2-3b,OpenRouter,Cloudflare,,,state,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,5,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| llama-3.2-3b,OpenRouter,Cloudflare,,,conflict,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,full,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |
| llama-3.2-3b,OpenRouter,Cloudflare,,,constraint,dual-instrumented task battery; esoteric contamination register,single-token answer,correctness,8,0|1|2|3|5|8|13|full,6,,true,The Evidence-Saturation Point of Large Language Models,Table 4,https://papers.ssrn.com/sol3/papers.cfm?abstract_id=7038059,https://github.com/hstre/evidence-k,Hanns-Steffen Rentschler,pilot, | |