モデル情報

Tiel-Coder-35B-A3B-GGUF-MTP

peculiar-ragdoll · Tiel-Coder-35B-A3B-GGUF-MTP

*Tiel* is the fast coder of the arsenal. At 4-bit quantization and 23 GB it fixes real codebase issues at the rate (and speed, with the right GPU) of Opus 4.6 medium, while holding the best multi-turn conversation of any local model we have measured. It is also cheerfully bad at trivia. > **This is Ornith-1.5-35B-A3B re-quantized > dynamically with our own imatrix and carrying the Sharp chat template** > inside the GGUF. Find the non-MTP GGUFs here, and the non-MTP MLX > version here. > **The uncensored and improved CyberTiel is out now!** Get it here. **Where it stands.** On 25 SWE-bench-Live problems Tiel fixes 12 — the same as Opus 4.6 (medium), four more than Ornith-1.5 itself, three more than Nail, and four more than Sonnet 5 (medium). Among models of its own class it is first; the ones ahead are dense 27Bs and Opus 5. Its time per attempt is also steadier than Nail's: an 8.6 minute median against 7.2, but a 12.3 minute mean against 15.7, because it lacks Nail's tail of expensive attempts.

翻訳待ちのため原文を表示しています。

コード生成license:mit
コンテキスト長非公開
入力 → 出力テキスト → テキスト
モデルウェイトオープンウェイト
情報更新日時2026/11/12 20:16:15
OVERVIEW

モデル紹介

*Tiel* is the fast coder of the arsenal. At 4-bit quantization and 23 GB it fixes real codebase issues at the rate (and speed, with the right GPU) of Opus 4.6 medium, while holding the best multi-turn conversation of any local model we have measured. It is also cheerfully bad at trivia. > **This is Ornith-1.5-35B-A3B re-quantized > dynamically with our own imatrix and carrying the Sharp chat template** > inside the GGUF. Find the non-MTP GGUFs here, and the non-MTP MLX > version here. > **The uncensored and improved CyberTiel is out now!** Get it here. **Where it stands.** On 25 SWE-bench-Live problems Tiel fixes 12 — the same as Opus 4.6 (medium), four more than Ornith-1.5 itself, three more than Nail, and four more than Sonnet 5 (medium). Among models of its own class it is first; the ones ahead are dense 27Bs and Opus 5. Its time per attempt is also steadier than Nail's: an 8.6 minute median against 7.2, but a 12.3 minute mean against 15.7, because it lacks Nail's tail of expensive attempts.

以下はAIが出典をもとに整理した要約で、当サイトの実測ではありません。原資料と実際の用途で検証してください。

主な特徴

  • Optimized for fast, practical codebase problem-solving at 4-bit quantization (23 GB size)
  • Supports multi-turn conversations better than other local models in its class
  • Uses the Sharp chat template and custom dynamic imatrix re-quantization of Ornith-1.5-35B-A3B

制約と注意点

  • Cheerfully bad at trivia tasks
  • Lower mean time per attempt (12.3 min) than Nail's (15.7 min) but higher median (8.6 vs 7.2 min), indicating less consistent speed under load
  • Performance lags behind dense 27B models and Opus 5 on SWE-bench-Live
PRICING & COST

出典別料金と費用分析

料金はチャネル・地域・設定によって異なります。各チャネルの料金は個別に表示し、提供元の統一公式料金とは見なしません。

出典・チャネル入力出力課金単位収集日時
確認可能な料金がありません。出典を確認してください。

月額の基本トークン費用を試算

—

試算はキャッシュ割引、段階料金、マルチモーダル追加料金、検索、税金、追加推論出力を含みません。

INDEPENDENT BENCHMARKS

公開評価と実行設定

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

収録済みの全ランキング →

よくある質問

What does 'MTP' stand for in the model name?

MTP is not explicitly defined in the source material; it appears as a naming suffix distinguishing this variant from non-MTP GGUF/MLX versions.

How does Tiel compare to Ornith-1.5-35B-A3B?

Tiel is a dynamically re-quantized version of Ornith-1.5-35B-A3B using a custom imatrix and embedding the Sharp chat template — it solves 4 more SWE-bench-Live problems than Ornith-1.5 itself.

Is this model uncensored?

The base Tiel-Coder-35B-A3B-GGUF-MTP is not stated to be uncensored; the announcement notes that 'the uncensored and improved CyberTiel is out now', implying this version is distinct and not uncensored.

オープンウェイトなら無料で商用利用できますか?

必ずしもそうではありません。バージョン別のライセンス、商用条件、利用方針を確認してください。自前運用には計算資源と保守費用も必要です。

これはリアルタイムの請求料金ですか?

収集日時付きの料金スナップショットです。最新のバージョン、地域、割引、課金方式は各チャネルで確認してください。

出典と追跡記録

许可证:mit。模型介绍与第三方评测均以原来源的具体版本和方法为准。