Seed 2.0 Pro

Seed 2.0 Pro is designed for long-chain reasoning and complex workflows that require sustained, multi-step problem solving. Its reported results include 98.3 on AIME 2025 and 97.3 on HMMT for mathematics, plus 74.0 on FrontierSci-olympiad for science reasoning, indicating especially strong quantitative and STEM performance.

The model is also a strong fit for software engineering and tool-oriented tasks, with reported scores of 76.5 on SWE-Bench Verified and 87.8 on LiveCodeBench v6. A reported 256k context window and support for function calling suit long-context agent workflows, including multi-step orchestration and adjustable thinking modes. The source also describes multimodal understanding, including images and video, for applications that need more than text-only analysis.

Merge Gatewaybytedance/dola-seed-2.0-proseed

Open provider page Provider docs

Quick Info

Provider: Merge Gateway
Model key: bytedance/dola-seed-2.0-pro
Release date: Mar 28, 2026
Last updated: Mar 28, 2026
Input modalities
Output modalities
Capabilities

Cost

Input token cost: $0.50
Output token cost: $3.00

Limits

Output tokens: 32,768 tokens
Context window: 131,072 tokens

Transparent token rates

Compare Seed 2.0 Pro pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about Seed 2.0 Pro

Merge Gateway

CoverageAnalysis

The LLM-Stats scorecard page explicitly names Seed 2.0 Pro and aggregates its reported benchmark scores, attributing them to seed.bytedance.com (ByteDance's first-party Seed team domain). Reported figures include AIME 2025 at 0.98/1, AIME 2026 at 0.94/1, and GPQA at 0.89/1 (Diamond methodology), with additional coverag As a third-party leaderboard aggregator, LLM-Stats does not introduce new technical events, API changes, or model updates for Seed 2.0 Pro; it functions as a snapshot dashboard that recycles first-party benchmark data. The subject rules require exact-version evidence, and this page satisfies that requirement by explici