Celeris 1 is a diffusion language model built for ultra-low-latency classification, extraction, judging, query rewriting, and other short structured responses.
Added Jul 25, 2026
Context Window
8.2K
Max Output
8.2K
Input Price (Auto)
$2.00/1M
Output Price (Auto)
$6.00/1M
Benchmarks
Benchmarks
Performance metrics and benchmarks
Sourced from Artificial Analysis.
Intelligence Index
6.3
Coding Index
14.4
Agentic work
AutomationBench-AA
Workflow automation with guardrail penalties
0.4%
Better than 8% of models compared
AA-Briefcase
Agentic knowledge work (Elo)
149 Elo
Better than 11% of models compared
GDPval-AA v2
Economically valuable tasks (Elo)
288 Elo
Better than 19% of models compared
Document reasoning
GDP.pdf
Professional PDF reasoning: all-pass rate
1.2%
Better than 9% of models compared
AA-LCR v1.1
Long context reasoning with updated grading
38.0%
Better than 36% of models compared
Reasoning
HLE
Humanity's Last Exam
6.8%
Better than 42% of models compared
Coding
Terminal-Bench v4.0
Practical coding and terminal tasks
0.0%
Better than 17% of models compared
SciCode
Python programming for scientific computing
21.6%
Better than 3% of models compared
Knowledge
MMLU-Pro
Professional and academic subject knowledge
78.0%
Better than 59% of models compared
Legacy benchmarks
GPQA Diamond (legacy)
Graduate-level scientific reasoning
63.1%
Better than 38% of models compared
AA-LCR (unversioned / legacy)
Long context reasoning evaluation
38.0%
Better than 36% of models compared
Last updated Sep 24, 2026
Artificial AnalysisProviders
Auto routing is available for this model. Explicit provider selection is not available.
Loading provider options…
Related text models
Compare Celeris 1 with similar models from the same provider or model family.
AionLabs: Aion 3.5
aion-labs/aion-3.5A GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.
AionLabs: Aion 3.5 Mini
aion-labs/aion-3.5-miniA GLM-family collaborative generation model tuned for immersive roleplay and storytelling, with stronger narrative structure, tension, conflict, and nuanced mature themes.
Qwen3.8 Max Prime
qwen/qwen3.8-max-primeQwen3.8 Max Prime is a higher-throughput variant of Qwen3.8 Max for coding, complex analysis, and long-running agent workflows. It accepts text, image, and video input, supports tool calling and structured output, and has a 1M-token context window. Reasoning is always enabled.
Space Bunny Alpha
stealth/space-bunny-alphaSpace Bunny Alpha is an anonymous stealth preview model for coding and multimodal tasks. It accepts text, images, and video, supports tool calling and structured output, and always reasons with adjustable effort across a 1M-token context window.
Solar Mini 4
upstage/solar-mini4Upstage's compact 35B-parameter mixture-of-experts model with 3B active parameters and a 524K context window. Built for fast, cost-efficient agentic tasks, with strong Korean and English support.
Solar Mini 4 Thinking
upstage/solar-mini4:thinkingSolar Mini 4 with reasoning enabled for agentic tasks and harder analysis across a 524K-token context window.