Entity · benchmark
AIME 2026
benchmarkactive
aime-2026-5418d929·6 events·first seen Aliases: AIME 2026, AIME 2024
Co-occurring entities
Lightning OPD 2.0LiveCodeBenchKlear-Reasoner-8B-SFTIFBenchBehavioral-SafetyBenchBrowseCompClawBenchBFCL-V3Mach-Mind-4-FlashMulti-Teacher On-Policy DistillationHybrid Median-length Policy Optimizationon-policy distillationWeak-to-Strong Generalization via Direct On-Policy DistillationQwen3-1.7BReinforcement Learning with Verifiable RewardsDAPOGRPOSTAREAIME 2025AceReason-14B
More like this (12)
Recent events (6)
Lightning OPD 2.0 mitigates style bias in cross-teacher on-policy distillation for reasoning models
A new arXiv preprint introduces Lightning OPD 2.0, a method for on-policy distillation (OPD) that addresses style bias when the SFT data generator and distillation teacher are different models. The approach uses rollout-level cross-fitting to estimate and subtract a 'style residual' from teacher-reference disagreement before constructing token-level updates. Starting from Klear-Reasoner-8B-SFT, the method achieves 82.4% on AIME 2024 and 63.0% on LiveCodeBench v5, outperforming the original Lightning OPD in cross-teacher settings. The work relaxes a key practical constraint in distillation pipelines by decoupling SFT data generation from the distillation teacher.
Mach-Mind-4-Flash: 35B MoE agentic model matching 100B-class performance via post-training optimization
Mach-Mind-4-Flash is a 35B-parameter Mixture-of-Experts model with only 3B activated parameters that achieves performance comparable to 100B-class models through post-training techniques alone. The pipeline combines a unified RL/OPD training infrastructure with multi-teacher scheduling, parallel domain-specific RL experts fused via Multi-Teacher On-Policy Distillation (MOPD), and Hybrid Median-length Policy Optimization (HMPO) which compresses reasoning chains 19-46% with minimal accuracy loss. Benchmark results include 92.70 on AIME'26, 82.82 on IFBench, and 75.80 on BFCL-v4, claiming to lead or match models 10-30x its activated size at a fraction of inference cost. The work is notable for demonstrating that post-training optimization can close large gaps in activated parameter count for agentic tasks.
Direct On-Policy Distillation transfers RL policy shifts from weak to strong models
Researchers propose Direct-OPD (Direct On-Policy Distillation), a method for transferring the policy shift induced by reinforcement learning on a small model to a larger target model, bypassing the need to run expensive RL rollouts on the stronger model. The approach uses the log-ratio between a post-RL teacher and its pre-RL reference as a dense implicit reward signal applied to the student's own on-policy states. Empirically, Direct-OPD improves Qwen3-1.7B from 48.3% to 62.4% on AIME 2024 in 4 hours on 8 A100 GPUs, outperforming step-matched direct RL. The method addresses a key scalability bottleneck in post-training as frontier models grow larger.
STARE: Token-level advantage reweighting to prevent entropy collapse in GRPO-style RL training
Researchers introduce STARE, a method addressing policy entropy collapse in GRPO-style reinforcement learning from verifiable rewards (RLVR) for LLM post-training. Through first-order gradient analysis, they identify a token-level credit assignment mismatch and propose selectively reweighting advantages for entropy-critical tokens using batch-internal surprisal quantiles plus a closed-loop entropy gate. Evaluated across 1.5B–32B models on short/long chain-of-thought and multi-turn tool use tasks, STARE outperforms DAPO and other baselines by 4–8% on AIME24/25 while sustaining stable training over thousands of steps.
ReasonAlloc: Hierarchical KV Cache Budget Allocation for Long-CoT Reasoning Models
ReasonAlloc is a training-free framework that reframes decoding-time KV cache compression as a hierarchical budget allocation problem, operating at both layer-wise (offline) and head-wise (online) levels. The method identifies an architecture-driven pattern called the 'Reasoning Wave' to guide layer preallocation, then dynamically reallocates to information-rich heads during decoding. Evaluated on MATH-500 and AIME 2024 using DeepSeek-R1-Distill and AceReason models, it outperforms uniform-budget baselines (R-KV, SnapKV, Pyramid-RKV) especially at small budgets of 128–512 tokens, with negligible overhead.
Z.ai's GLM-5.1 Open-Weights Model Targets Multi-Hour Agentic Coding Tasks with Iterative Self-Evaluation
Z.ai released GLM-5.1, a 754B parameter mixture-of-experts open-weights model optimized for long-running agentic coding tasks, capable of cycling through planning, execution, and strategy revision hundreds of times over sessions lasting up to eight hours. The model achieves top open-weights scores on the Artificial Analysis Intelligence Index and third place on Arena's Code leaderboard, while leading SWE-Bench Pro in Z.ai's own evaluations at 58.4 percent. Weights are available on HuggingFace under MIT license, with API pricing roughly 40 percent higher than its predecessor but still below comparable proprietary models. No technical report has been published, leaving architecture and training details undisclosed.