What's new
now data Recomputed benchmark-weighted quality scores now data Updated speed measurements now data Synced Chatbot Arena benchmark track now data Validated official pricing snapshots now data Pulled latest OpenRouter price index 1h jobs Jobs market snapshot refreshed 25 May digest OpenGuardrails: An Open-Source Context-Aware AI Guardrails Platform 25 May digest Published the 2026-05-25 daily digest 25 May digest Workbench Launches Open Source BullMQ Dashboard For Node Backends 24 May digest SpecBench Tests Reward Hacking In Long Horizon Coding Agents now data Recomputed benchmark-weighted quality scores now data Updated speed measurements now data Synced Chatbot Arena benchmark track now data Validated official pricing snapshots now data Pulled latest OpenRouter price index 1h jobs Jobs market snapshot refreshed 25 May digest OpenGuardrails: An Open-Source Context-Aware AI Guardrails Platform 25 May digest Published the 2026-05-25 daily digest 25 May digest Workbench Launches Open Source BullMQ Dashboard For Node Backends 24 May digest SpecBench Tests Reward Hacking In Long Horizon Coding Agents
Historical snapshot

This evidence was captured on 12 June 2026. It is provisional and is not a current ranking.

Mistral

Mistral top model

Track only if a new Mistral release lands near the current frontier.

0 Floor score

confidence: low

0 Floor Capability

Finished benchmark shape.

0 Operating Envelope

No current comparable operating row.

0 Frontier Reality

Watchlist only

0 Research Trust

No current comparable research-trust evidence.

Current signal

Watchlist only

Evidence completeness

15%

1 score dimensions represented.

DSWE

Not reported

Keep watch for a DeepSWE row.

Calculation

Score components

0 Accuracy + grounding

Correct, grounded answers on ordinary verifiable tasks.

0 Instruction discipline

Follows exact user constraints and output formats.

0 Honesty + reliability

Avoids unsupported confidence and remains stable across runs.

0 Reasoning floor

Handles non-flashy reasoning without brittle failures.

0 Agent/coding

Completes tool, coding, and agent-style tasks when evidence exists.

0 Operating envelope

Cost, speed, token burn, verbosity, and friction.

Evidence

Sources used

  • watchlist

Caveat

What to watch

Not ranked in the first table because there is no recent frontier-level Mistral signal comparable to the current leaders.

Meaning

How to read the score

Each score is a 0-100 saturation scale. A score of 100 means that benchmark lane is complete for this model. Floor Capability is the finished benchmark shape; Operating Envelope, Frontier Reality, and Research Trust are provisional tabs that need prompt-pack backfill. Confidence is evidence quality, not another model grade. Evidence % is source coverage, not capability.

Citations

Benchmark sources