whatmodel.sh / Doubao Seed 2.1 Pro (260628)

Doubao Seed 2.1 Pro (260628) is a closed-weights language model from ByteDance Seed, released June 2026. Licence: proprietary.

Catalogue identity notes

Exact Volcengine Ark model id 'doubao-seed-2-1-pro-260628' (model list, observed 2026-11-12), the Seed2.1 Pro checkpoint released with the Seed2.1 family (ByteDance Seed blog dated 2026-06-23: 'Doubao and Volcano Engine users can now start to access Doubao Seed 2.1'). Distinct from seed-2-1-pro-preview (Arena 'seed-2.1-pro-preview', a pre-release Code Arena checkpoint whose own note says it is not asserted to equal the released checkpoints) and from doubao-seed-2-1-pro-260915 (later checkpoint, 1024K context). Matches Epoch version 'doubao-seed-2.1-pro-260628' (dots for hyphens), Epoch release date 2026-06-23.

Model information sources

Release date ↗

70.4±1.8#79 of 743 LLMsfrom 4 benchmark results across 1 source

Score within the LMArena (text) comparison set only.

$ how it compares

The LLMs ranked either side of Doubao Seed 2.1 Pro (260628), out of 897 with a score, then more from ByteDance Seed.

#modelscorewhereRAMreleased
83Qwen3.7 Plus Alibaba70±1api $0.70/M2026-06
84DeepSeek V4 Pro (Preview, Apr 2026) DeepSeek70±1local~632GB2026-04
85Doubao Seed 2.1 Pro (260628) ByteDance Seed70±2api $1.78/M2026-06
86Doubao Seed 2.1 Turbo (260628) ByteDance Seed70±2api $0.89/M2026-06
87Claude Opus 4.5 Anthropic70±1api $10.00/M2025-11
more from ByteDance Seed
90Doubao Seed 2.0 Pro (260215) ByteDance Seed70±1api $0.95/M2026-02
100Doubao Seed 2.0 Lite (260428) ByteDance Seed69±1api $0.20/M2026-04
220Seed-OSS 36B Instruct ByteDance Seed62±1local~23GB2025-08

see all LLMs · find the ones that fit your machine

$ what it's good at

math▁▁▃▅█▆▅█▄▁▁▁ #60 of 440 * · 9.3% of its score · 1 test

Beats LongCat 2.0 (1.6T-A48B), Claude Sonnet 4.6, Gemini 3 Flash (Preview) here, though they rank higher overall. Trails A.X K2 (688B-A33B), DeepSeek V4 Flash (Preview, Apr 2026), Doubao Seed 2.1 Turbo (260628) here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
OTIS mock AIME 2024–202595.6% ✎▃▃▄▃▃▃▇▆█▄▃▆ beats 163/207very hard · avg 7.4%▇▇▇▇ 9.3%sharp · noisy
other tests3 not in a task · 89% of its score
testresultvs the others in its rundifficultycountswhy
Chess Puzzles (Epoch)21.0% ✎█▁▁▁▁▂▄▂▃▃▂▁ beats 87/157very hard · avg 5.6%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 42%sharp · noisy
GPQA Diamond89.3% ✎▁▁▁▁▂▃▅▄▄▆█▃ beats 171/221hard · avg 39%▇▇▇▇▇▇▇▇▇▇▇▇▇ 33%sharp
SimpleQA Verified44.7% ✎▁▂▂▂▅█▄▃▂··▁ beats 47/91very hard · avg 7.4%▇▇▇▇▇▇ 15%at its level

$ where to run it

Through an API

api · Volcengine Ark · $1.78/M · $0.8906 in / $4.45301 out per 1M tokens · 256k context · docs.volcengine.com ↗ · price sourceUSD per 1M tokens from models.dev (CNY list price converted at 6.737012 CNY/USD, 2026-08-26); cache read excluded. Volcengine's own pricing page (observed 2026-11-12) lists doubao-seed-2.1-pro 'input [0, 1024]K: 6.00 / cache hit 1.20 / output 30.00' CNY per 1M tokens, consistent.

$ how the score was worked out · Epoch AI — internal evaluatio… 99%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Epoch AI — internal evaluations4CC BY 4.0
prior · pull toward the middle
weighted average= 70.4

$ improve this page

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json