swarmagent.dev
GitHub
DocsModelsResourcesPricingNewsContactGitHub

Fireworks AI model

Kimi K2.7 Code

Kimi K2.7 Code is a coding-focused agentic model built upon Kimi K2.6. With substantial improvements on real-world long-horizon coding tasks, it strengthens end-to-end task completion across complex software engineering workflows while improving token efficiency, reducing thinking-token usage by approximately 30% compared with Kimi K2.6.

Provider
Fireworks AI
Availability
Available
Model ID
kimi-k2p7-code
Snapshot
2026-10-07
Evaluation Verdict

Moonshot AI Kimi K2.7 Code provides an extraordinary 2M token context window paired with specialized code reasoning (61.5% SWE-bench Verified, 81.2% AIME, 1370 Coding Elo).

Kimi K2.7 Code Benchmarks & Agentic Accuracy

Verified primary research & developer preference metrics

Primary Sources Verified
SWE-bench Verified
61.5%
Long-Context ReasoningMoonshot AI Report ↗
LiveCodeBench
56.8%
ThinkingLiveCodeBench ↗
LMSYS Coding Elo
1370Elo
Coding ArenaLMSYS Arena ↗
AIME 2024 / 2025
81.2%
ThinkingMoonshot AI Report ↗
Pricing ($ / 1M tokens)
$0.6/ $2
$0.15 cached readStandard API Pricing

What This Means in Real Software Development

  • Massive 2M token context window allows ingesting entire monorepos without middle-context truncation.
  • 61.5% on SWE-bench Verified proves strong multi-file debugging and refactoring competence.
  • 81.2% on AIME 2024 demonstrates superior multi-step logic and mathematical reasoning.
  • Priced at $0.60/M input on Fireworks AI, offering exceptional value for large-context code tasks.

Swarm Multi-Agent Topology

Tier 2: Long-Context Code SpecialistRole: Coder Subagent

Best Paired With: Gemini 3.8 Flash for subagent execution

Exceptional capability in large repositories where cross-module dependency trees must fit in context.

Git Worktree Isolation Advantage:

Executes full-repo refactors inside dedicated Git worktrees.

What Developers & Practitioners Are Saying

Real-world field observations, production strengths, and developer consensus

Field Consensus

"Moonshot AI's specialized coding model, leveraging deep context retrieval and long-horizon coherence."

Verified Real-World Strengths

  • 61.5% on SWE-bench Verified
  • Excellent long-context repository recall
  • High fidelity on complex database migrations and backend schemas

Production Caveats & Gaps

  • Less widespread global edge hosting compared to US-based providers
Benchmark Reality Check

What the benchmark tests: Tests long-horizon issue resolution and extensive repository context retention.

Production caveat: Use prompt caching when analyzing large codebases.

Primary Sources & Methodology Citations

Kimi K2.7 Code model overview

Kimi K2.7 Code is listed in the public Swarm catalog through Fireworks AI. This page summarizes the verified catalog fields used by Swarm without claiming real-time access for every provider account.

Provider
Fireworks AI
Family
Kimi
Context window
262.1K tokens
Thinking support
Not supported
Availability status
Available
Catalog ID
fireworks/kimi-k2p7-code
  • Image input
  • Text input
  • Text output
  • Tools
  • Fine-tuning

Run Kimi K2.7 Code in Swarm

Verified CLI workflow & multi-agent architecture

Quickstart Guide →
bash — swarm run
$swarm ctl model set --provider fireworks --model fireworks/kimi-k2p7-code

# Or switch interactively inside the terminal TUI:

>/models

Subagent Role Matrix

Assign Kimi K2.7 Code to specialized system roles in ~/.swarm/config.json:

  • Primary / Planner: Orchestrates full task DAGs and multi-stage pipelines.
  • Coder Subagent: Generates code and unit tests inside private Git worktrees.
  • Finder Subagent: Performs read-only code search, symbol discovery, and mapping.
System Agent Roles →Why Swarm Architecture →
← All Fireworks AI models

Public catalog snapshot. Catalog availability is not a real-time provider status or an access guarantee.