AI Models · Open-Source vs Paid

Step 3.7 Flash Open vs Codestral 2508 (batch) Paid

Step 3.7 Flash vs Codestral 2508 (batch) compared — price per token, context window, multimodality, openness and which to choose. Can the open-source model replace the paid one? Full 2026 breakdown.

Codestral 2508 (batch) — full profile ›

Prices & specs refreshed from live data · olud.ai

Open-model prices = cheapest provider via OpenRouter; official maker rates may be higher.

Step 3.7 FlashOpenStepfun

$0.2 /M input
$1.15 /M outputNo per-token fees if you self-host

TypeOpen-weight

Context window262K tokens

MultimodalYes

Self-hostYes

Codestral 2508 (batch)PaidMistral AI

$0.15 /M input
$0.45 /M outputManaged API (no infra to run)

TypeProprietary

Context window256K tokens

MultimodalNo

Self-hostNo

Choose Step 3.7 Flash if you want to self-host, keep your data private and skip per-token fees — it's open-weight and runs on your own hardware. Choose Codestral 2508 (batch) if you want frontier capability through a managed API with zero infrastructure to run.

Step 3.7 Flash vs Codestral 2508 (batch) specs

Spec	Step 3.7 Flash	Codestral 2508 (batch)	Winner
Maker	Stepfun	Mistral AI	–
Type	Open-weight	Proprietary	Step 3.7 Flash
Context window	262K tokens	256K tokens	Step 3.7 Flash
Input price	$0.2/M · free self-host	$0.15/M	Codestral 2508 (batch)
Output price	$1.15/M · free self-host	$0.45/M	Codestral 2508 (batch)
Vision / multimodal	Yes	No	Step 3.7 Flash
Tool / function calling	Yes	Yes	= Tie
Self-hostable	Yes	No (API only)	Step 3.7 Flash
License	Open weights	Proprietary	Step 3.7 Flash

Price gap & when to choose each

2.6×cheaper per output token

Codestral 2508 (batch) is ~2.6× cheaper than Step 3.7 Flash on output tokens ($0.45 vs $1.15 per M tokens).

Choose Step 3.7 Flash if…

You want to self-host or run on your cloud
You need the longer 262K context window
You prioritize data privacy & control
You are building open or reproducible AI

Choose Codestral 2508 (batch) if…

You want frontier performance through a managed API
You want the lower output price
You value reliability & ecosystem
You don't want to manage infrastructure

Feature comparison

Capability	Step 3.7 Flash	Codestral 2508 (batch)
Open weights (downloadable)	✓	✗
Self-hostable	✓	✗
Runs fully offline	✓	✗
Vision / multimodal	✓	✗
Tool / function calling	✓	✓
1M+ context window	✗	✗

Benchmarks: Step 3.7 Flash vs Codestral 2508 (batch)

Independent benchmark scores measured by Artificial Analysis. Higher is better (except latency).

Step 3.7 Flash

Codestral 2508 (batch)

Intelligence index

19.5

–

Coding index

39.6

–

GPQA

80.9%

–

Humanity's Last Exam

21.4%

–

Long Context Reasoning

73.7%

–

SciCode

43.9%

–

IFBench

67.3%

–

τ²-Bench

98.5%

–

τ-Bench Banking

12%

–

Terminal-Bench

39.3%

–

Terminal-Bench Hard

35.6%

–

Speed

132.8 tok/s

–

Latency

1.59s

–

Intelligence per $

44.5

–

Benchmark data by Artificial Analysis.

How Step 3.7 Flash and Codestral 2508 (batch) score

🏆 Best value & openness: Step 3.7 Flash (4.7 vs 3.0 / 5)

Criterion	Step 3.7 Flash	Codestral 2508 (batch)
Cost-efficiency	4.5	5.0
Context window	4.0	4.0
Openness	5.0	1.5
Self-hosting	5.0	1.0
Multimodality	5.0	3.5

Scores come from live data — output price (cost), context length, open vs closed weights (openness & self-hosting) and vision/tool support (multimodality). Raw task quality isn't scored here; it depends on your benchmark — see the verdict.

What each model is

Step 3.7 Flash Open

Stepfun · Open-weight

Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...

Codestral 2508 (batch) Paid

Mistral AI · Proprietary

Mistral's cutting-edge language model for coding released end of July 2025. Codestral specializes in low-latency, high-frequency tasks such as fill-in-the-middle (FIM), code correction and test generation. [Blog Post](https://mistral.ai/ne

Other models in these families

These variants are tracked but not compared here — one page per family keeps the comparison readable.

Other variants tracked

Step 3.5 Flash

Frequently asked questions

Is Step 3.7 Flash as good as Codestral 2508 (batch)?

Step 3.7 Flash is open-weight and competitive on many tasks, but Codestral 2508 (batch) may still lead on the hardest reasoning and agentic work. The gap keeps narrowing — benchmark both on your actual use case before deciding.

Can I run Step 3.7 Flash locally?

Yes. Step 3.7 Flash has open weights, so you can self-host it on your own GPUs or run it via a low-cost API. Codestral 2508 (batch) is API-only and cannot be self-hosted.

Step 3.7 Flash vs Codestral 2508 (batch) — which should I pick in 2026?

Choose Step 3.7 Flash if you want to self-host, keep your data private and skip per-token fees — it's open-weight and runs on your own hardware. Choose Codestral 2508 (batch) if you want frontier capability through a managed API with zero infrastructure to run.

People also compare

Hy3 vs Codestral 2508 (batch)GLM 5.3 vs Codestral 2508 (batch)DeepSeek V4.1 Flash vs Codestral 2508 (batch)MiMo-V2.5-Pro vs Codestral 2508 (batch)Nemotron 3 Ultra vs Codestral 2508 (batch)Kimi K3 (batch) vs Codestral 2508 (batch)

Explore more open-source AI

Browse the full open-source model leaderboard, thousands of tools and live benchmarks — all in one place.

Open the leaderboard →