Sanity Harness scores Kimi K2.6, Opus 4.7

// 45d agoBENCHMARK RESULT

Sanity Harness scores Kimi K2.6, Opus 4.7

Sanity Harness’s latest leaderboard adds 145 results across older and newer runs, including fresh tests of Kimi K2.6-Code-Preview, Opus 4.7, GLM 5.1, and Minimax M2.7. The author’s main takeaway is that Opus 4.7 is a real step up, Kimi K2.6 still looks early, and GLM 5.1 lands near the top of the open-weight pack.

// ANALYSIS

This is less a model launch than a reality check: the frontier still looks meaningfully ahead, but the margins inside the top tier are getting clearer and more interesting.

–Opus 4.7 appears to be the strongest signal in the batch, which matters because many recent “upgrades” have been mostly marketing.
–Kimi K2.6-Code-Preview is promising, but the post itself treats it as premature evidence rather than a final verdict.
–GLM 5.1 seems to be the best open-weight showing here, while Minimax M2.7 sits in the useful middle tier for price and local deployment.
–ForgeCode’s strong Minimax result is interesting, but the author says the tool is buggy and too workflow-specific to recommend broadly yet.
–Sanity Harness’s value is the methodology: sandboxed runs, Docker validation, and weighted scoring make the leaderboard more credible than a single-model demo.
–For coding-agent buyers, this reinforces a familiar split: frontier models still buy reliability, while open-weight options buy cost control and deployability.

// TAGS

sanity-harnessbenchmarkai-codingagentcliopen-weights

DISCOVERED

45d ago

2026-04-17

PUBLISHED

45d ago

2026-04-17

RELEVANCE

9/ 10

AUTHOR

lemon07r

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

OPEN SOURCE1h ago

parakeet.cpp delivers 2x faster local ASR

Developed by the LocalAI team, parakeet.cpp is a dependency-free C++17 inference engine for NVIDIA's NeMo Parakeet ASR models that runs up to 2x faster than standard baselines. By leveraging the ggml library to eliminate Python runtime dependencies, it enables highly portable offline speech recognition across CPUs and multiple GPU backends.

MODEL1h ago

NVIDIA has launched Nemotron 3 Ultra, a massive 550-billion parameter hybrid Mamba-Transformer model that is five times faster and 30% cheaper than comparable open models.

NVIDIA has unveiled Nemotron 3 Ultra, the flagship model in its new Nemotron 3 family of foundation models built specifically for complex reasoning, planning, and multi-step agentic workflows. Featuring 550 billion total parameters, the model is powered by a hybrid Mamba-Transformer Mixture-of-Experts (MoE) architecture that utilizes LatentMoE token compression and NVFP4 training to activate only 55 billion parameters per token. This efficient design enables throughputs exceeding 300 tokens per second, a 1-million token context window, and up to 5x faster inference at 30% lower cost compared to existing open-weights models in its class.

VIDEO1h ago

Socket Firewall blocks dependency malware in real time by intercepting package manager network requests at install time.

Socket Firewall is a free command-line interface (CLI) tool designed to proactively protect developer environments from malicious software supply chain attacks. By acting as a local network proxy or command prefix, the tool intercepts package manager network requests (supporting npm, pip, and cargo) during installation, evaluating dependencies in real-time against Socket's security API. This allows it to automatically block known malware, flag suspicious packages with risky capabilities (like unexpected telemetry, network, or filesystem access), and enforce safety policies without disrupting developer workflows.