Models

Choose a harness, then copy a model name from its tab. Use the same model string in the CLI, TypeScript, and Python.

Setting	CLI → managed SDK
Harness	`-a` → `agents[].name`
Model	`-m` → `agents[].model_name`
Reasoning effort	`--effort` → `agents[].reasoning_effort`

Every evaluation arm requires a model. The default model below is the Agent SDK’s choice when its model is omitted. Default effort is Evolve’s configured setting when effort is omitted.

Model and effort reference

Reasoning options are checked against the exact model’s provider documentation and the harness’s native controls. Models with a thinking toggle or a token budget are labeled in the table. off is the CLI and SDK spelling for thinking disabled and thinking for enabled; none is a native spelling shown in some tables and is not accepted. A model outside a harness’s roster runs only through the Agent SDK; managed evaluations refuse it.

This page records native support. Some Evolve effort mappings still need correction before all of these options work through the CLI and SDK.

Harness: claudeDefault model: opus or claude-opus-5-5Default effort: high

Model	Model string	Native reasoning efforts
Claude Fable 5.1	`fable`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Opus 5.5	`opus`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Sonnet 5.5	`sonnet`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Haiku 4.5	`haiku`	Thinking budget; no effort levels
GLM 5.3	`glm-5.3`	`low`, `high`, `max`
GLM 5.3 Prime	`glm-5.3-prime`	`low`, `high`, `max`
GLM 5.3 Flash	`glm-5.3-flash`	`low`, `high`, `max`
GLM 5.3 FlashX	`glm-5.3-flashx`	`low`, `high`, `max`
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	`low`, `high`, `max`
DeepSeek V4.1 Flash	`fireworks/deepseek-v4.1-flash`	`low`, `high`, `max`

Sources

Claude Code · Claude effort · GLM 5.3 · Fireworks model controls

Other accepted model spellings

Model string	Also accepted
`fable`	`claude-fable-5-1`
`opus`	`claude-opus-5-5`
`sonnet`	`claude-sonnet-5-5`
`haiku`	`claude-haiku-4-5-20251001`

Harness: codexDefault model: gpt-6.1-solDefault effort: high

Model	Model string	Native reasoning efforts
GPT-6 Astra	`gpt-6-astra`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-6.1 Sol	`gpt-6.1-sol`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-5.6 Terra	`gpt-5.6-terra`	`none`, `low`, `medium`, `high`, `xhigh`, `max`
GPT-6 Luna	`gpt-6-luna`	`none`, `low`, `medium`, `high`, `xhigh`, `max`
GPT-5.3 Codex	`gpt-5.3-codex`	`low`, `medium`, `high`, `xhigh`

Sources

GPT-6 Astra · GPT-6.1 Sol · GPT-6 Luna · GPT-5.6 Terra · GPT-5.3 Codex

Harness: qwenDefault model: qwen3.8-maxDefault effort: thinking

Model	Model string	Native reasoning efforts
Qwen 3.8 Max	`qwen3.8-max`	`off`, `low`, `medium`, `xhigh`
Qwen 3.8 Max Prime	`qwen3.8-max-prime`	`minimal`, `low`, `medium`, `high`, `xhigh` (always thinking; cannot be turned off)
Qwen 3.7 Plus	`qwen3.7-plus`	`off`, `thinking`
Qwen 3.8 Flash	`qwen3.8-flash`	`off`, `low`, `medium`, `xhigh`

Sources

Qwen Code settings · Qwen model controls · OpenRouter reasoning

Harness: kimiDefault model: kimi-k3 or moonshot/kimi-k3Default effort: thinking

Model	Model string	Native reasoning efforts
Kimi K3	`kimi-k3`	`low`, `high`, `max`
Kimi K2.7 Code	`kimi-k2.7-code`	Always thinking; no effort levels
Kimi K3 Raptor	`kimi-k3-raptor`	`off`, `low`, `high`, `max`

Sources

Kimi model controls · Kimi reasoning · Fireworks reasoning

Other accepted model spellings

Model string	Also accepted
`kimi-k3`	`moonshot/kimi-k3`
`kimi-k2.7-code`	`moonshot/kimi-k2.7-code`

Harness: opencodeDefault model: openrouter/anthropic/claude-opus-5.5Default effort: high

Model	Model string	Native reasoning efforts
Claude Fable 5.1	`openrouter/anthropic/claude-fable-5.1`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Opus 5.5	`openrouter/anthropic/claude-opus-5.5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Sonnet 5.5	`openrouter/anthropic/claude-sonnet-5.5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Haiku 4.5	`openrouter/anthropic/claude-haiku-4.5`	Thinking budget; no effort levels
GPT-6 Astra	`openrouter/openai/gpt-6-astra`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-6.1 Sol	`openrouter/openai/gpt-6.1-sol`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-5.6 Terra	`openrouter/openai/gpt-5.6-terra`	`none`, `low`, `medium`, `high`, `xhigh`, `max`
GPT-6 Luna	`openrouter/openai/gpt-6-luna`	`none`, `low`, `medium`, `high`, `xhigh`, `max`
Gemini 3.8 Flash	`openrouter/google/gemini-3.8-flash`	`low`, `medium`, `high`
Qwen 3.8 Max	`openrouter/qwen/qwen3.8-max-0902`	`minimal`, `low`, `medium`, `high`, `xhigh`
Qwen 3.8 Max Prime	`openrouter/qwen/qwen3.8-max-prime`	`minimal`, `low`, `medium`, `high`, `xhigh`
Kimi K3	`openrouter/moonshotai/kimi-k3`	`none`, `low`, `high`, `max`
GLM 5.3	`openrouter/z-ai/glm-5.3`	`low`, `high`, `max`
GLM 5.3 Prime	`openrouter/z-ai/glm-5.3-prime`	`low`, `high`, `max`
GLM 5.3 Flash	`openrouter/z-ai/glm-5.3-flash`	`low`, `high`, `max`
GLM 5.3 FlashX	`openrouter/z-ai/glm-5.3-flashx`	`low`, `high`, `max`
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	`none`, `low`, `high`, `max`
DeepSeek V4.1 Flash	`fireworks/deepseek-v4.1-flash`	`none`, `low`, `high`, `max`

Sources

OpenCode models · OpenRouter reasoning · Fireworks model controls

Harness: droidDefault model: claude-opus-5-5Default effort: high

Model	Model string	Native reasoning efforts
Claude Fable 5.1	`claude-fable-5.1`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Opus 5.5	`claude-opus-5-5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Sonnet 5.5	`claude-sonnet-5-5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Haiku 4.5	`claude-haiku-4-5`	`off`, `low`, `medium`, `high` (thinking budgets)
GPT-5.6 Terra	`gpt-5.6-terra`	`none`, `low`, `medium`, `high`, `xhigh`, `max`
Gemini 3.8 Flash	`gemini-3.8-flash`	`low`, `medium`, `high`
Qwen 3.8 Max	`qwen3.8-max`	`low`, `medium`, `xhigh`
Qwen 3.8 Max Prime	`qwen3.8-max-prime`	`low`, `medium`, `high`, `xhigh`
Kimi K3	`kimi-k3`	`low`, `high`, `max`
GLM 5.3	`glm-5.3`	`low`, `high`, `max`
GLM 5.3 Prime	`glm-5.3-prime`	`low`, `high`, `max`
GLM 5.3 Flash	`glm-5.3-flash`	`low`, `high`, `max`
GLM 5.3 FlashX	`glm-5.3-flashx`	`low`, `high`, `max`
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	`off`, `low`, `high`, `max`
DeepSeek V4.1 Flash	`fireworks/deepseek-v4.1-flash`	`off`, `low`, `high`, `max`

Sources

Factory models · Claude effort · Gemini thinking · Fireworks model controls

Other accepted model spellings

Model string	Also accepted
`claude-fable-5.1`	`claude-fable-5-1`
`claude-haiku-4-5`	`claude-haiku-4-5-20251001`

Harness: piDefault model: openrouter/anthropic/claude-opus-5.5Default effort: high

Model	Model string	Native reasoning efforts
Claude Fable 5.1	`openrouter/anthropic/claude-fable-5.1`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Opus 5.5	`openrouter/anthropic/claude-opus-5.5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Sonnet 5.5	`openrouter/anthropic/claude-sonnet-5.5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Haiku 4.5	`openrouter/anthropic/claude-haiku-4.5`	`off`, `minimal`, `low`, `medium`, `high` (thinking budgets)
GPT-6 Astra	`openrouter/openai/gpt-6-astra`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-6.1 Sol	`openrouter/openai/gpt-6.1-sol`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-5.6 Terra	`openrouter/openai/gpt-5.6-terra`	`off`, `low`, `medium`, `high`, `xhigh`, `max`
GPT-6 Luna	`openrouter/openai/gpt-6-luna`	`off`, `low`, `medium`, `high`, `xhigh`, `max`
Gemini 3.8 Flash	`openrouter/google/gemini-3.8-flash`	`low`, `medium`, `high`
Qwen 3.8 Max	`openrouter/qwen/qwen3.8-max-0902`	`minimal`, `low`, `medium`, `high`, `xhigh`
Qwen 3.8 Max Prime	`openrouter/qwen/qwen3.8-max-prime`	`minimal`, `low`, `medium`, `high`, `xhigh`
Kimi K3	`openrouter/moonshotai/kimi-k3`	`off`, `low`, `high`, `max`
GLM 5.3	`openrouter/z-ai/glm-5.3`	`low`, `high`, `max`
GLM 5.3 Prime	`openrouter/z-ai/glm-5.3-prime`	`low`, `high`, `max`
GLM 5.3 Flash	`openrouter/z-ai/glm-5.3-flash`	`low`, `high`, `max`
GLM 5.3 FlashX	`openrouter/z-ai/glm-5.3-flashx`	`low`, `high`, `max`
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	`off`, `low`, `high`, `max`

Sources

Pi model configuration · OpenRouter reasoning

Harness: prime-agentDefault model: openrouter/anthropic/claude-opus-5.5Default effort: high

Model	Model string	Native reasoning efforts
Claude Fable 5.1	`openrouter/anthropic/claude-fable-5.1`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Opus 5.5	`openrouter/anthropic/claude-opus-5.5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Sonnet 5.5	`openrouter/anthropic/claude-sonnet-5.5`	`low`, `medium`, `high`, `xhigh`, `max`
Claude Haiku 4.5	`openrouter/anthropic/claude-haiku-4.5`	`off`, `high` (thinking off/on; no effort levels)
GPT-6 Astra	`openrouter/openai/gpt-6-astra`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-6.1 Sol	`openrouter/openai/gpt-6.1-sol`	`low`, `medium`, `high`, `xhigh`, `max`
GPT-5.6 Terra	`openrouter/openai/gpt-5.6-terra`	`off`, `low`, `medium`, `high`, `xhigh`, `max`
GPT-6 Luna	`openrouter/openai/gpt-6-luna`	`off`, `low`, `medium`, `high`, `xhigh`, `max`
Gemini 3.8 Flash	`openrouter/google/gemini-3.8-flash`	`low`, `medium`, `high`
Qwen 3.8 Max	`openrouter/qwen/qwen3.8-max-0902`	`minimal`, `low`, `medium`, `high`, `xhigh`
Qwen 3.8 Max Prime	`openrouter/qwen/qwen3.8-max-prime`	`minimal`, `low`, `medium`, `high`, `xhigh`
Kimi K3	`openrouter/moonshotai/kimi-k3`	`off`, `low`, `high`, `max`
GLM 5.3	`openrouter/z-ai/glm-5.3`	`low`, `high`, `max`
GLM 5.3 Prime	`openrouter/z-ai/glm-5.3-prime`	`low`, `high`, `max`
GLM 5.3 Flash	`openrouter/z-ai/glm-5.3-flash`	`low`, `high`, `max`
GLM 5.3 FlashX	`openrouter/z-ai/glm-5.3-flashx`	`low`, `high`, `max`
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	`off`, `low`, `high`, `max`

Sources

Prime Agent · OpenRouter reasoning

Harness: dshDefault model: openrouter/deepseek/deepseek-v4.1-flashDefault effort: high

Model	Model string	Native reasoning efforts
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	`off`, `low`, `high`, `max`
DeepSeek V4.1 Flash	`fireworks/deepseek-v4.1-flash`	`off`, `low`, `high`, `max`
DeepSeek V4 Pro	`openrouter/deepseek/deepseek-v4-pro-0813`	`off`, `low`, `high`, `max`

Sources

DeepSeek Harness · OpenRouter reasoning · Fireworks reasoning

Harness: zcodeDefault model: openrouter/z-ai/glm-5.3Default effort: high

Model	Model string	Native reasoning efforts
GLM 5.3	`openrouter/z-ai/glm-5.3`	`low`, `high`, `max`
GLM 5.3 Prime	`openrouter/z-ai/glm-5.3-prime`	`low`, `high`, `max`
GLM 5.3 Flash	`openrouter/z-ai/glm-5.3-flash`	`low`, `high`, `max`
GLM 5.3 FlashX	`openrouter/z-ai/glm-5.3-flashx`	`low`, `high`, `max`
GLM 5.3	`fireworks/glm-5.3`	`low`, `high`, `max`
GLM 5.3 Flash	`fireworks/glm-5.3-flash`	`low`, `high`, `max`

Sources

Z Code · GLM 5.3 · GLM 5.3 Flash · Fireworks model controls

Harness: antigravityDefault model: gemini-3.8-flash or vertex_ai/gemini-3.8-flashDefault effort: high

Model	Model string	Native reasoning efforts
Gemini 3.8 Flash	`gemini-3.8-flash`	`low`, `medium`, `high`
Gemini 3.5 Flash-Lite	`gemini-3.5-flash-lite`	`low`, `medium`, `high`
Gemini 3.1 Pro Preview	`gemini-3.1-pro-preview`	`low`, `medium`, `high`

Sources

Antigravity CLI · Gemini thinking

Other accepted model spellings

Model string	Also accepted
`gemini-3.8-flash`	`vertex_ai/gemini-3.8-flash`
`gemini-3.5-flash-lite`	`vertex_ai/gemini-3.5-flash-lite`
`gemini-3.1-pro-preview`	`vertex_ai/gemini-3.1-pro-preview`

Harness: mini-swe-agentDefault model: claude-sonnet-5-5Default effort: none

mini-SWE-agent runs one bash tool through LiteLLM’s chat completions and has no effort setting of its own. Every model goes through OpenRouter (or the Evolve gateway, which routes there). It takes no MCP servers and no skills, and runs at most 250 steps.

Model	Model string	Native reasoning efforts
Claude Sonnet 5.5	`claude-sonnet-5-5`	None: mini sends no reasoning setting, so the model uses its own default
Claude Opus 5.5	`claude-opus-5-5`	None: mini sends no reasoning setting, so the model uses its own default
Claude Haiku 4.5	`claude-haiku-4-5`	None: mini sends no reasoning setting, so the model uses its own default
GPT-6.1 Sol	`gpt-6.1-sol`	None: mini sends no reasoning setting, so the model uses its own default
GPT-6 Luna	`gpt-6-luna`	None: mini sends no reasoning setting, so the model uses its own default
DeepSeek V4.1 Flash	`openrouter/deepseek/deepseek-v4.1-flash`	None: mini sends no reasoning setting, so the model uses its own default
GLM 5.3	`openrouter/z-ai/glm-5.3`	None: mini sends no reasoning setting, so the model uses its own default
GLM 5.3 Prime	`openrouter/z-ai/glm-5.3-prime`	None: mini sends no reasoning setting, so the model uses its own default
Kimi K3	`openrouter/moonshotai/kimi-k3`	None: mini sends no reasoning setting, so the model uses its own default

Sources

mini-SWE-agent · OpenRouter models

gemini is retired. Use antigravity for new runs. Records of earlier Gemini CLI runs remain readable.

Use your selection

This example runs Codex with GPT-6 Luna at high effort.

evolve run \
  -d deepswe@1.1 \
  -i anko-default-function-arguments \
  -a codex \
  -m gpt-6-luna \
  --effort high \
  --max-trial-spend 1 \
  --max-retries 0 \
  --watch

import { jobs } from "@evolvingmachines/evolve";

const job = await jobs().start({
  datasets: [{
    name: "deepswe",
    version: "1.1",
    task_names: ["anko-default-function-arguments"],
  }],
  agents: [{
    name: "codex",
    model_name: "gpt-6-luna",
    reasoning_effort: "high",
  }],
  max_trial_spend_usd: 1,
  retry: { max_retries: 0 },
});

import asyncio
from evolve import jobs


async def main():
    job = await jobs().start(
        datasets=[{
            "name": "deepswe",
            "version": "1.1",
            "task_names": ["anko-default-function-arguments"],
        }],
        agents=[{
            "name": "codex",
            "model_name": "gpt-6-luna",
            "reasoning_effort": "high",
        }],
        max_trial_spend_usd=1,
        retry={"max_retries": 0},
    )
    print(job.id)


asyncio.run(main())

Repeat -m to compare models on one harness. --effort applies to every arm. To compare different efforts, set reasoning_effort separately on each arm in a job config or the SDK.

Using the Agent SDK instead of managed evaluations

The Agent SDK uses model and TypeScript reasoningEffort / Python reasoning_effort in its agent configuration. It can omit the model; managed evaluation arms cannot.The tables above describe Evolve model access. Direct provider credentials have their own available models. On Claude Code and Droid, the glm-5.3…, openrouter/… and fireworks/… rows, and Droid’s qwen3.8-max… rows, are gateway routes, and the fireworks/ options on OpenCode, DeepSeek Harness, and Z Code require Evolve model access.See Configure an agent for SDK examples.

Read the current model list

curl -sS https://dashboard.evolvingmachines.ai/api/meta

No authentication is required. agents[].models lists model aliases and IDs. Metadata reports the API’s accepted inputs and configured defaults.The same document is available through meta() in both SDKs.

Model access and spend

An EVOLVE_API_KEY supplies platform model access. --max-trial-spend caps agent model spend per trial attempt. Verifier judges and trace analysis have separate budgets. See job costs.

Provider keys and metering

With your own supported provider key, model requests bill that provider account. Evolve gateway metering and the trial cap still apply. Job creation still requires a positive platform credit balance.Requests that bypass the gateway with separate credentials are not metered by Evolve. A task’s no-network agent policy restricts model access to the platform gateway.

Claude and GPT model behaviour

On Claude Code, Codex and Droid, the Claude and GPT models above (and their anthropic/ and openai/ spellings) differ from the model maker’s own API in three ways:

Claude’s server-side web search tool does not run.
Claude’s thinking blocks come back without signatures.
A Claude thinking budget is sent as an effort level: minimal below 1,024 tokens, low from 1,024, medium from 5,000, high from 10,000.

Requests made with your own Anthropic or OpenAI key, and Reward Kit judges, are not affected.

Analysis and check models

Analysis and task checks choose their agent and model separately from the evaluation arms.

Operation	Agent / model options
`evolve analyze`	`-a` / `-m`
`evolve check`	`-a` / `-m`
Analysis during `evolve run`	`--analyze-agent` / `--analyze-model`

Read the current analysis and check defaults before choosing an agent, model, or effort. A named model must be supported by the selected harness.

evolve analyze --show-defaults

evolve check --show-defaults

Pass -a and -m to override these defaults for one analysis or check.

Configure a run

Model comparisons, effort, concurrency, and spend controls.

Start here

Run evaluations

Task format

Validate your work

Read results

Work together

Model and effort reference

Use your selection

Model access and spend

Analysis and check models

Configure a run